fix: batch queries on recall

2026-01-13 11:54:54 +01:00 · 2026-01-13 11:40:48 +01:00
390 changed files with 11888 additions and 60144 deletions
@@ -153,15 +153,8 @@ jobs:
    - name: Build docs
      run: npm run build --workspace=hindsight-docs

-  test-rust-cli:
+  build-rust-cli:
    runs-on: ubuntu-latest
-    env:
-      HINDSIGHT_API_LLM_PROVIDER: groq
-      HINDSIGHT_API_LLM_API_KEY: ${{ secrets.GROQ_API_KEY }}
-      HINDSIGHT_API_LLM_MODEL: openai/gpt-oss-20b
-      HINDSIGHT_API_URL: http://localhost:8888
-      GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      UV_INDEX: pytorch=https://download.pytorch.org/whl/cpu

    steps:
    - uses: actions/checkout@v4
@@ -178,10 +171,6 @@ jobs:
          hindsight-cli/target
        key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}

-    - name: Run unit tests
-      working-directory: hindsight-cli
-      run: cargo test
-
    - name: Build CLI
      working-directory: hindsight-cli
      run: cargo build --release
@@ -193,6 +182,29 @@ jobs:
        path: hindsight-cli/target/release/hindsight
        retention-days: 1

+  test-rust-cli:
+    runs-on: ubuntu-latest
+    needs: build-rust-cli
+    env:
+      HINDSIGHT_API_LLM_PROVIDER: groq
+      HINDSIGHT_API_LLM_API_KEY: ${{ secrets.GROQ_API_KEY }}
+      HINDSIGHT_API_LLM_MODEL: openai/gpt-oss-20b
+      HINDSIGHT_API_URL: http://localhost:8888
+      GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      UV_INDEX: pytorch=https://download.pytorch.org/whl/cpu
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: Download CLI artifact
+      uses: actions/download-artifact@v4
+      with:
+        name: hindsight-cli
+        path: /tmp/cli
+
+    - name: Make CLI executable
+      run: chmod +x /tmp/cli/hindsight
+
    - name: Install uv
      uses: astral-sh/setup-uv@v5
      with:
@@ -239,7 +251,7 @@ jobs:

    - name: Run CLI smoke test
      run: |
-        HINDSIGHT_CLI=hindsight-cli/target/release/hindsight ./hindsight-cli/smoke-test.sh
+        HINDSIGHT_CLI=/tmp/cli/hindsight ./hindsight-cli/smoke-test.sh

    - name: Show API server logs
      if: always()
@@ -765,7 +777,7 @@ jobs:

  test-doc-examples:
    runs-on: ubuntu-latest
-    needs: test-rust-cli
+    needs: build-rust-cli
    env:
      HINDSIGHT_API_LLM_PROVIDER: groq
      HINDSIGHT_API_LLM_API_KEY: ${{ secrets.GROQ_API_KEY }}
@@ -875,66 +887,6 @@ jobs:
        echo "=== API Server Logs ==="
        cat /tmp/api-server.log || echo "No API server log found"

-  test-upgrade:
-    runs-on: ubuntu-latest
-    env:
-      HINDSIGHT_API_LLM_PROVIDER: groq
-      HINDSIGHT_API_LLM_API_KEY: ${{ secrets.GROQ_API_KEY }}
-      HINDSIGHT_API_LLM_MODEL: openai/gpt-oss-20b
-      GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      UV_INDEX: pytorch=https://download.pytorch.org/whl/cpu
-
-    steps:
-    - uses: actions/checkout@v4
-      with:
-        fetch-depth: 0  # Full history needed for git clone of tags
-
-    - name: Fetch tags
-      run: git fetch --tags
-
-    - name: Install uv
-      uses: astral-sh/setup-uv@v5
-      with:
-        enable-cache: true
-        prune-cache: false
-
-    - name: Set up Python
-      uses: actions/setup-python@v5
-      with:
-        python-version-file: ".python-version"
-
-    - name: Cache HuggingFace models
-      uses: actions/cache@v4
-      with:
-        path: ~/.cache/huggingface
-        key: ${{ runner.os }}-huggingface-${{ hashFiles('hindsight-api/pyproject.toml') }}
-        restore-keys: |
-          ${{ runner.os }}-huggingface-
-
-    - name: Install hindsight-dev dependencies
-      working-directory: ./hindsight-dev
-      run: uv sync --frozen --extra test --index-strategy unsafe-best-match
-
-    - name: Install current hindsight-api
-      working-directory: ./hindsight-api
-      run: uv sync --frozen --index-strategy unsafe-best-match
-
-    - name: Pre-download models
-      working-directory: ./hindsight-api
-      run: |
-        uv run python -c "
-        from sentence_transformers import SentenceTransformer, CrossEncoder
-        print('Downloading embedding model...')
-        SentenceTransformer('BAAI/bge-small-en-v1.5')
-        print('Downloading cross-encoder model...')
-        CrossEncoder('cross-encoder/ms-marco-MiniLM-L-6-v2')
-        print('Models downloaded successfully')
-        "
-
-    - name: Run upgrade tests
-      working-directory: ./hindsight-dev
-      run: uv run pytest upgrade_tests/ -v --tb=short
-
  verify-generated-files:
    runs-on: ubuntu-latest
    env:
@@ -29,7 +29,7 @@ nltk_data/

 # Monitoring stack (Prometheus/Grafana binaries and data)
 .monitoring/
-.pgbouncer/
+.pgbouncer

 # Large benchmark datasets (will be downloaded automatically)
 **/longmemeval_s_cleaned.json
@@ -1,3 +1,153 @@
 # AGENTS.md

-See [CLAUDE.md](./CLAUDE.md) for project documentation and coding conventions.
+This document captures architectural decisions and coding conventions for the Hindsight project.
+
+## Documentation
+
+- **Main documentation**: [hindsight-docs/docs/developer/](./hindsight-docs/docs/developer/)
+- **Use case patterns**: [hindsight-docs/docs/cookbook/](./hindsight-docs/docs/cookbook/)
+- **API reference**: Auto-generated from OpenAPI spec
+
+## Project Structure
+
+```
+hindsight/              # Python package for embedded usage
+hindsight-api/          # FastAPI server (core memory engine)
+hindsight-cli/          # Rust CLI client
+hindsight-embed/        # Embedded CLI (no server needed)
+hindsight-control-plane/ # Next.js admin UI
+hindsight-docs/         # Docusaurus documentation site
+hindsight-dev/          # Development tools and benchmarks
+hindsight-integrations/ # Framework integrations (LangChain, etc.)
+hindsight-clients/      # Generated API clients (Python, TypeScript, Rust)
+```
+
+## Core Concepts
+
+### Memory Banks
+- Each bank is an isolated memory store (like a "brain" for one user/agent)
+- Banks contain: memory units (facts), entities, documents, entity links
+- Banks have a **disposition** (personality traits) and **background** (context)
+- Bank isolation is strict - no cross-bank data leakage
+
+### Memory Types
+- **World facts**: General knowledge ("The sky is blue")
+- **Experience facts**: Personal experiences ("I visited Paris in 2023")
+- **Opinion facts**: Beliefs with confidence scores ("Paris is beautiful" - 0.9 confidence)
+
+### Operations
+- **Retain**: Store new memories (extracts facts, entities, relationships)
+- **Recall**: Retrieve memories (semantic, BM25, graph, temporal search)
+- **Reflect**: Deep analysis to form new insights/opinions
+
+## API Design Decisions
+
+### Single Bank Per Request
+- All API endpoints (`recall`, `reflect`, `retain`) operate on a single bank
+- Multi-bank queries are the **client/agent's responsibility** to orchestrate
+- This keeps the API simple and the isolation model clear
+
+### Disposition Traits (3-trait system)
+- **Skepticism** (1-5): How skeptical vs trusting when forming opinions
+- **Literalism** (1-5): How literally to interpret information
+- **Empathy** (1-5): How much to consider emotional context
+- These influence the `reflect` operation, not `recall`
+- Background info also only affects `reflect` (opinion formation)
+
+## Multi-Bank Architecture Patterns
+
+See [hindsight-docs/docs/cookbook/](./hindsight-docs/docs/cookbook/) for detailed guides:
+
+- **Per-User Memory**: One bank per user, simplest pattern
+- **Support Agent + Shared Knowledge**: User bank + shared docs bank, client orchestrates
+
+## Developer Guide
+
+### Running the API Server
+
+```bash
+# From project root
+./scripts/dev/start-api.sh
+
+# With options
+./scripts/dev/start-api.sh --reload --port 8888 --log-level debug
+```
+
+### Running Tests
+
+```bash
+# API tests
+cd hindsight-api
+uv run pytest tests/
+
+# Specific test
+uv run pytest tests/test_http_api_integration.py -v
+```
+
+### Generating OpenAPI Spec
+
+After changing API endpoints, regenerate the OpenAPI spec and docs:
+
+```bash
+./scripts/generate-openapi.sh
+```
+
+This will:
+1. Generate `openapi.json` at project root
+2. Copy to `hindsight-docs/openapi.json`
+3. Regenerate API reference documentation
+
+### Generating API Clients
+
+After updating the OpenAPI spec, regenerate all clients:
+
+```bash
+./scripts/generate-clients.sh
+```
+
+This generates:
+- **Rust client**: `hindsight-clients/rust/` (via progenitor in build.rs)
+- **Python client**: `hindsight-clients/python/` (via openapi-generator Docker)
+- **TypeScript client**: `hindsight-clients/typescript/` (via @hey-api/openapi-ts)
+
+Note: The maintained wrapper `hindsight_client.py` and `README.md` are preserved during regeneration.
+
+### Running the Documentation Site
+
+```bash
+./scripts/dev/start-docs.sh
+```
+
+### Running the Control Plane
+
+```bash
+./scripts/dev/start-control-plane.sh
+```
+
+## Code Style
+
+### Python (hindsight-api)
+- Use `uv` for package management
+- Async throughout (asyncpg, async FastAPI endpoints)
+- Pydantic models for request/response validation
+- No py files at project root - maintain clean directory structure
+
+### TypeScript (control-plane, clients)
+- Next.js with App Router for control plane
+- Tailwind CSS with shadcn/ui components
+
+### Rust (CLI)
+- Async with tokio
+- reqwest for HTTP client
+- progenitor for API client generation
+
+## Database
+
+- PostgreSQL with pgvector extension
+- Schema managed via Alembic migrations in `hindsight-api/alembic/`, db migrations happen during api startup, no manual commands
+- Key tables: `banks`, `memory_units`, `documents`, `entities`, `entity_links`
+
+# Branding
+## Colors
+- Primary: gradient from #0074d9 to #009296
+
@@ -7,7 +7,8 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 Hindsight is an agent memory system that provides long-term memory for AI agents using biomimetic data structures. Memories are organized as:
 - **World facts**: General knowledge ("The sky is blue")
 - **Experience facts**: Personal experiences ("I visited Paris in 2023")
- **Mental models**: Consolidated knowledge synthesized from facts ("User prefers functional programming patterns")
+- **Opinion facts**: Beliefs with confidence scores ("Paris is beautiful" - 0.9 confidence)
+- **Observations**: Complex mental models derived from reflection

 ## Development Commands

@@ -100,7 +101,7 @@ cd hindsight-control-plane && npm run dev
 Main operations:
 - **Retain**: Store memories, extracts facts/entities/relationships
 - **Recall**: Retrieve memories via 4 parallel strategies (semantic, BM25, graph, temporal) + reranking
- **Reflect**: Disposition-aware reasoning using memories and mental models.
+- **Reflect**: Deep analysis forming new opinions/observations (disposition-aware)

 ### Database
 PostgreSQL with pgvector. Schema managed via Alembic migrations in `hindsight-api/hindsight_api/alembic/`. Migrations run automatically on API startup.
@@ -173,63 +174,12 @@ This runs the same checks as the pre-commit hook (Ruff for Python, ESLint/Pretti
 - Multi-bank queries are client responsibility to orchestrate
 - Disposition traits only affect reflect, not recall

-### Control Plane API Routes
-
-When adding or modifying parameters in the dataplane API (hindsight-api), you must also update the control plane routes that proxy to it:
-
-1. **API Routes** (`hindsight-control-plane/src/app/api/`):
-   - `recall/route.ts` - proxies to `/v1/default/banks/{bank_id}/memories/recall`
-   - `reflect/route.ts` - proxies to `/v1/default/banks/{bank_id}/reflect`
-   - `memories/retain/route.ts` - proxies to `/v1/default/banks/{bank_id}/memories/retain`
-   - Other routes follow the same pattern
-
-2. **Client types** (`hindsight-control-plane/src/lib/api.ts`):
-   - Update the TypeScript type definitions for `recall()`, `reflect()`, `retain()` etc.
-
-3. **Checklist when adding new API parameters**:
-   - Add parameter extraction in the route handler (destructure from `body`)
-   - Pass the parameter to the SDK call
-   - Update the client type definition in `lib/api.ts`
-   - Update any UI components that need to use the new parameter
-
 ### Python Style
 - Python 3.11+, type hints required
 - Async throughout (asyncpg, async FastAPI)
 - Pydantic models for request/response
 - Ruff for linting (line-length 120)
 - No Python files at project root - maintain clean directory structure
- **Never use multi-item tuple return values** - prefer dataclass or Pydantic model for structured returns
-
-### Type Safety with Pydantic Models
-**NEVER use raw `dict` types for structured data.** Always use Pydantic models:
- Use Pydantic `BaseModel` for all data structures passed between functions
- Add `@field_validator` for type coercion (e.g., ensuring datetimes are timezone-aware)
- Avoid `dict.get()` patterns - use typed model attributes instead
- Parse external data (JSON, API responses) into Pydantic models at the boundary
- This catches type errors at parse time, not deep in business logic
-
-```python
-# BAD - error-prone dict access
-def process(data: dict) -> str:
-    return data.get("name", "")  # No validation, silent failures
-
-# GOOD - typed and validated
-class UserData(BaseModel):
-    name: str
-    created_at: datetime
-
-    @field_validator("created_at", mode="before")
-    @classmethod
-    def ensure_tz_aware(cls, v):
-        if isinstance(v, str):
-            v = datetime.fromisoformat(v.replace("Z", "+00:00"))
-        if v.tzinfo is None:
-            return v.replace(tzinfo=timezone.utc)
-        return v
-
-def process(data: UserData) -> str:
-    return data.name  # Type-safe, validated at construction
-```

 ### TypeScript Style
 - Next.js App Router for control plane
@@ -5,7 +5,7 @@
 [Documentation](https://hindsight.vectorize.io) • [Paper](https://arxiv.org/abs/2512.12818) • [Cookbook](https://hindsight.vectorize.io/cookbook) • [Hindsight Cloud](https://vectorize.io/hindsight/cloud)

 [![CI](https://github.com/vectorize-io/hindsight/actions/workflows/release.yml/badge.svg)](https://github.com/vectorize-io/hindsight/actions/workflows/release.yml)
-[![Slack Community](https://img.shields.io/badge/Slack-Join%20Community-4A154B?logo=slack)](https://join.slack.com/t/hindsight-space/shared_invite/zt-3nhbm4w29-LeSJ5Ixi6j8PdiYOCPlOgg)
+[![Slack Community](https://img.shields.io/badge/Slack-Join%20Community-4A154B?logo=slack)](https://join.slack.com/t/hindsight-space/shared_invite/zt-3klo21kua-VUCC_zHP5rIcXFB1_5yw6A)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 ![PyPI - Downloads](https://img.shields.io/pypi/dm/hindsight-api?label=PyPI)
 ![NPM Downloads](https://img.shields.io/npm/dm/%40vectorize-io%2Fhindsight-client?logoColor=orange&label=NPM&color=blue&link=https%3A%2F%2Fwww.npmjs.com%2Fpackage%2F%40vectorize-io%2Fhindsight-client)
@@ -242,7 +242,7 @@ client.reflect(bank_id="my-bank", query="What should I know about Alice?")
 - [CLI](https://hindsight.vectorize.io/sdks/cli)

 **Community:**
- [Slack](https://join.slack.com/t/hindsight-space/shared_invite/zt-3nhbm4w29-LeSJ5Ixi6j8PdiYOCPlOgg)
+- [Slack](https://join.slack.com/t/hindsight-space/shared_invite/zt-3klo21kua-VUCC_zHP5rIcXFB1_5yw6A)
 - [GitHub Issues](https://github.com/vectorize-io/hindsight/issues)

 ---
@@ -2,8 +2,8 @@ apiVersion: v2
 name: hindsight
 description: Hindsight helm chart
 type: application
-version: 0.3.0
-appVersion: "0.3.0"
+version: 0.2.1
+appVersion: "0.2.1"
 keywords:
  - ai
  - memory
@@ -80,22 +80,6 @@ Control plane selector labels
 app.kubernetes.io/component: control-plane
 {{- end }}

-{{/*
-Worker labels
-*/}}
-{{- define "hindsight.worker.labels" -}}
-{{ include "hindsight.labels" . }}
-app.kubernetes.io/component: worker
-{{- end }}
-
-{{/*
-Worker selector labels
-*/}}
-{{- define "hindsight.worker.selectorLabels" -}}
-{{ include "hindsight.selectorLabels" . }}
-app.kubernetes.io/component: worker
-{{- end }}
-
 {{/*
 Create the name of the service account to use
 */}}
@@ -55,11 +55,6 @@ spec:
        {{- end }}
        - name: HINDSIGHT_API_DATABASE_URL
          value: {{ include "hindsight.databaseUrl" . | quote }}
-        {{- /* Disable internal worker when dedicated workers are enabled */}}
-        {{- if .Values.worker.enabled }}
-        - name: HINDSIGHT_API_WORKER_ENABLED
-          value: "false"
-        {{- end }}
        {{- range $key, $value := .Values.api.env }}
        - name: {{ $key }}
          value: {{ $value | quote }}
@@ -1,25 +0,0 @@
-{{- if .Values.worker.enabled }}
-apiVersion: v1
-kind: Service
-metadata:
-  name: {{ include "hindsight.fullname" . }}-worker
-  labels:
-    {{- include "hindsight.worker.labels" . | nindent 4 }}
-  {{- if .Values.podAnnotations }}
-  annotations:
-    {{- /* Common Prometheus annotations for metrics scraping */}}
-    prometheus.io/scrape: "true"
-    prometheus.io/port: {{ .Values.worker.service.port | quote }}
-    prometheus.io/path: "/metrics"
-  {{- end }}
-spec:
-  # Headless service for StatefulSet (enables stable DNS names like worker-0.worker.namespace)
-  clusterIP: None
-  ports:
-  - port: {{ .Values.worker.service.port }}
-    targetPort: {{ .Values.worker.service.targetPort }}
-    protocol: TCP
-    name: http
-  selector:
-    {{- include "hindsight.worker.selectorLabels" . | nindent 4 }}
-{{- end }}
@@ -1,110 +0,0 @@
-{{- if .Values.worker.enabled }}
-apiVersion: apps/v1
-kind: StatefulSet
-metadata:
-  name: {{ include "hindsight.fullname" . }}-worker
-  labels:
-    {{- include "hindsight.worker.labels" . | nindent 4 }}
-spec:
-  serviceName: {{ include "hindsight.fullname" . }}-worker
-  replicas: {{ .Values.worker.replicaCount }}
-  selector:
-    matchLabels:
-      {{- include "hindsight.worker.selectorLabels" . | nindent 6 }}
-  template:
-    metadata:
-      annotations:
-        {{- if not .Values.existingSecret }}
-        checksum/secret: {{ include (print $.Template.BasePath "/secret.yaml") . | sha256sum }}
-        {{- end }}
-        {{- with .Values.podAnnotations }}
-        {{- toYaml . | nindent 8 }}
-        {{- end }}
-      labels:
-        {{- include "hindsight.worker.selectorLabels" . | nindent 8 }}
-    spec:
-      {{- if .Values.serviceAccount.create }}
-      serviceAccountName: {{ include "hindsight.serviceAccountName" . }}
-      {{- end }}
-      securityContext:
-        {{- toYaml .Values.podSecurityContext | nindent 8 }}
-      containers:
-      - name: worker
-        securityContext:
-          {{- toYaml .Values.securityContext | nindent 10 }}
-        image: "{{ .Values.worker.image.repository }}:{{ .Values.worker.image.tag | default .Values.version }}"
-        imagePullPolicy: {{ .Values.worker.image.pullPolicy }}
-        command: ["hindsight-worker"]
-        ports:
-        - name: http
-          containerPort: {{ .Values.worker.service.targetPort }}
-          protocol: TCP
-        {{- if .Values.existingSecret }}
-        envFrom:
-        - secretRef:
-            name: {{ .Values.existingSecret }}
-        {{- end }}
-        env:
-        {{- /* POSTGRES_PASSWORD must be defined before DATABASE_URL for $(VAR) interpolation */}}
-        {{- if not .Values.postgresql.enabled }}
-        - name: POSTGRES_PASSWORD
-          valueFrom:
-            secretKeyRef:
-              name: {{ include "hindsight.secretName" . }}
-              key: postgres-password
-        {{- end }}
-        - name: HINDSIGHT_API_DATABASE_URL
-          value: {{ include "hindsight.databaseUrl" . | quote }}
-        {{- /* Worker ID uses pod name (StatefulSet provides stable names like worker-0, worker-1) */}}
-        - name: HINDSIGHT_API_WORKER_ID
-          valueFrom:
-            fieldRef:
-              fieldPath: metadata.name
-        {{- /* Inherit LLM config from api.env */}}
-        {{- range $key, $value := .Values.api.env }}
-        - name: {{ $key }}
-          value: {{ $value | quote }}
-        {{- end }}
-        {{- /* Worker-specific env vars */}}
-        {{- range $key, $value := .Values.worker.env }}
-        - name: {{ $key }}
-          value: {{ $value | quote }}
-        {{- end }}
-        {{- /* Only use secrets when not using existingSecret */}}
-        {{- if not .Values.existingSecret }}
-        {{- /* Inherit secrets from api.secrets */}}
-        {{- range $key, $value := .Values.api.secrets }}
-        - name: {{ $key }}
-          valueFrom:
-            secretKeyRef:
-              name: {{ include "hindsight.secretName" $ }}
-              key: {{ $key }}
-        {{- end }}
-        {{- /* Worker-specific secrets (can override api.secrets) */}}
-        {{- range $key, $value := .Values.worker.secrets }}
-        - name: {{ $key }}
-          valueFrom:
-            secretKeyRef:
-              name: {{ include "hindsight.secretName" $ }}
-              key: {{ $key }}
-        {{- end }}
-        {{- end }}
-        livenessProbe:
-          {{- toYaml .Values.worker.livenessProbe | nindent 10 }}
-        readinessProbe:
-          {{- toYaml .Values.worker.readinessProbe | nindent 10 }}
-        resources:
-          {{- toYaml .Values.worker.resources | nindent 10 }}
-      {{- with .Values.nodeSelector }}
-      nodeSelector:
-        {{- toYaml . | nindent 8 }}
-      {{- end }}
-      {{- with .Values.affinity }}
-      affinity:
-        {{- toYaml . | nindent 8 }}
-      {{- end }}
-      {{- with .Values.tolerations }}
-      tolerations:
-        {{- toYaml . | nindent 8 }}
-      {{- end }}
-{{- end }}
@@ -67,63 +67,6 @@ api:
    # HINDSIGHT_API_LLM_API_KEY: "your-api-key"
    # HINDSIGHT_API_LLM_BASE_URL: "https://api.groq.com/openai/v1"

-# Worker settings (distributed task processing)
-# When enabled, dedicated worker pods process tasks and the API's internal worker is disabled
-worker:
-  enabled: false
-  replicaCount: 2
-  image:
-    repository: ghcr.io/vectorize-io/hindsight-api
-    pullPolicy: IfNotPresent
-    # tag defaults to .Values.version if not specified
-
-  service:
-    # Service for metrics scraping (headless for StatefulSet)
-    port: 8889
-    targetPort: 8889
-
-  # Resource limits and requests
-  resources:
-    limits:
-      cpu: 2000m
-      memory: 4Gi
-    requests:
-      cpu: 500m
-      memory: 1Gi
-
-  # Liveness and readiness probes
-  livenessProbe:
-    httpGet:
-      path: /health
-      port: 8889
-    initialDelaySeconds: 30
-    periodSeconds: 10
-    timeoutSeconds: 5
-    failureThreshold: 3
-
-  readinessProbe:
-    httpGet:
-      path: /health
-      port: 8889
-    initialDelaySeconds: 10
-    periodSeconds: 5
-    timeoutSeconds: 3
-    failureThreshold: 3
-
-  # Worker-specific environment variables
-  env:
-    # Poll interval in milliseconds (how often to check for new tasks)
-    HINDSIGHT_API_WORKER_POLL_INTERVAL_MS: "500"
-    # Number of tasks to claim per poll cycle
-    HINDSIGHT_API_WORKER_BATCH_SIZE: "10"
-    # Max retries before marking a task as failed
-    HINDSIGHT_API_WORKER_MAX_RETRIES: "3"
-    # HTTP port for metrics/health (matches service.targetPort)
-    HINDSIGHT_API_WORKER_HTTP_PORT: "8889"
-
-  # Secret environment variables (inherited from api.secrets if not specified)
-  secrets: {}
-
 # Image settings for control plane
 controlPlane:
  enabled: true
@@ -244,65 +244,6 @@ def run_db_migration(
    typer.echo("Database migrations completed successfully")


-async def _decommission_worker(db_url: str, worker_id: str, schema: str = "public") -> int:
-    """Release all tasks owned by a worker, setting them back to pending status."""
-    is_pg0, instance_name, _ = parse_pg0_url(db_url)
-    if is_pg0:
-        typer.echo(f"Starting embedded PostgreSQL (instance: {instance_name})...")
-    resolved_url = await resolve_database_url(db_url)
-
-    conn = await asyncpg.connect(resolved_url)
-    try:
-        table = _fq_table("async_operations", schema)
-        result = await conn.fetch(
-            f"""
-            UPDATE {table}
-            SET status = 'pending', worker_id = NULL, claimed_at = NULL, updated_at = now()
-            WHERE worker_id = $1 AND status = 'processing'
-            RETURNING operation_id
-            """,
-            worker_id,
-        )
-        return len(result)
-    finally:
-        await conn.close()
-
-
-@app.command(name="decommission-worker")
-def decommission_worker(
-    worker_id: str = typer.Argument(..., help="Worker ID to decommission"),
-    schema: str = typer.Option("public", "--schema", "-s", help="Database schema"),
-    yes: bool = typer.Option(False, "--yes", "-y", help="Skip confirmation prompt"),
-):
-    """Release all tasks owned by a worker (sets status back to pending).
-
-    Use this command when a worker has crashed or been removed without graceful shutdown.
-    All tasks that were being processed by the worker will be released back to the queue
-    so other workers can pick them up.
-    """
-    config = HindsightConfig.from_env()
-
-    if not config.database_url:
-        typer.echo("Error: Database URL not configured.", err=True)
-        typer.echo("Set HINDSIGHT_API_DATABASE_URL environment variable.", err=True)
-        raise typer.Exit(1)
-
-    if not yes:
-        typer.confirm(
-            f"This will release all tasks owned by worker '{worker_id}' back to pending. Continue?",
-            abort=True,
-        )
-
-    typer.echo(f"Decommissioning worker '{worker_id}' (schema: {schema})...")
-
-    count = asyncio.run(_decommission_worker(config.database_url, worker_id, schema))
-
-    if count > 0:
-        typer.echo(f"Released {count} task(s) from worker '{worker_id}'")
-    else:
-        typer.echo(f"No tasks found for worker '{worker_id}'")
-
-
 def main():
    app()

@@ -1,48 +0,0 @@
-"""add_tags_column
-
-Revision ID: g2a3b4c5d6e7
-Revises: f1a2b3c4d5e6
-Create Date: 2025-01-13
-
-Add tags column to memory_units and documents tables for visibility scoping.
-Tags enable filtering memories by scope (e.g., user IDs, session IDs) during recall/reflect.
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-# revision identifiers, used by Alembic.
-revision: str = "g2a3b4c5d6e7"
-down_revision: str | Sequence[str] | None = "f1a2b3c4d5e6"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (e.g., 'tenant_x.' or '' for public)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Add tags column to memory_units and documents tables."""
-    schema = _get_schema_prefix()
-
-    # Add tags column to memory_units table
-    op.execute(f"ALTER TABLE {schema}memory_units ADD COLUMN IF NOT EXISTS tags VARCHAR[] NOT NULL DEFAULT '{{}}'")
-
-    # Create GIN index for efficient array containment queries (tags && ARRAY['x'])
-    op.execute(f"CREATE INDEX IF NOT EXISTS idx_memory_units_tags ON {schema}memory_units USING GIN (tags)")
-
-    # Add tags column to documents table for document-level tags
-    op.execute(f"ALTER TABLE {schema}documents ADD COLUMN IF NOT EXISTS tags VARCHAR[] NOT NULL DEFAULT '{{}}'")
-
-
-def downgrade() -> None:
-    """Remove tags columns and index."""
-    schema = _get_schema_prefix()
-
-    op.execute(f"DROP INDEX IF EXISTS {schema}idx_memory_units_tags")
-    op.execute(f"ALTER TABLE {schema}memory_units DROP COLUMN IF EXISTS tags")
-    op.execute(f"ALTER TABLE {schema}documents DROP COLUMN IF EXISTS tags")
@@ -1,112 +0,0 @@
-"""mental_models_v4
-
-Revision ID: h3c4d5e6f7g8
-Revises: g2a3b4c5d6e7
-Create Date: 2026-01-08 00:00:00.000000
-
-This migration implements the v4 mental models system:
-1. Deletes existing observation memory_units (observations now in mental models)
-2. Adds mission column to banks (replacing background)
-3. Creates mental_models table with final schema
-
-Mental models can reference entities when an entity is "promoted" to a mental model.
-Summary content is stored as JSONB observations with per-observation fact attribution.
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-# revision identifiers, used by Alembic.
-revision: str = "h3c4d5e6f7g8"
-down_revision: str | Sequence[str] | None = "g2a3b4c5d6e7"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Apply mental models v4 changes."""
-    schema = _get_schema_prefix()
-
-    # Step 1: Delete observation memory_units (cascades to unit_entities links)
-    # Observations are now handled through mental models, not memory_units
-    op.execute(f"DELETE FROM {schema}memory_units WHERE fact_type = 'observation'")
-
-    # Step 2: Drop observation-specific index (if it exists)
-    op.execute(f"DROP INDEX IF EXISTS {schema}idx_memory_units_observation_date")
-
-    # Step 3: Add mission column to banks (replacing background)
-    op.execute(f"ALTER TABLE {schema}banks ADD COLUMN IF NOT EXISTS mission TEXT")
-
-    # Migrate: copy background to mission if background column exists
-    # Use DO block to check column existence first (idempotent for re-runs)
-    schema_name = context.config.get_main_option("target_schema") or "public"
-    op.execute(f"""
-        DO $$
-        BEGIN
-            IF EXISTS (
-                SELECT 1 FROM information_schema.columns
-                WHERE table_schema = '{schema_name}' AND table_name = 'banks' AND column_name = 'background'
-            ) THEN
-                UPDATE {schema}banks
-                SET mission = background
-                WHERE mission IS NULL;
-            END IF;
-        END $$;
-    """)
-
-    # Remove background column (replaced by mission)
-    op.execute(f"ALTER TABLE {schema}banks DROP COLUMN IF EXISTS background")
-
-    # Step 4: Create mental_models table with final v4 schema (if not exists)
-    op.execute(f"""
-        CREATE TABLE IF NOT EXISTS {schema}mental_models (
-            id VARCHAR(64) NOT NULL,
-            bank_id VARCHAR(64) NOT NULL,
-            subtype VARCHAR(32) NOT NULL,
-            name VARCHAR(256) NOT NULL,
-            description TEXT NOT NULL,
-            entity_id UUID,
-            observations JSONB DEFAULT '{{"observations": []}}'::jsonb,
-            links VARCHAR[],
-            tags VARCHAR[] DEFAULT '{{}}',
-            last_updated TIMESTAMP WITH TIME ZONE,
-            created_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT now(),
-            PRIMARY KEY (id, bank_id),
-            FOREIGN KEY (bank_id) REFERENCES {schema}banks(bank_id) ON DELETE CASCADE,
-            FOREIGN KEY (entity_id) REFERENCES {schema}entities(id) ON DELETE SET NULL,
-            CONSTRAINT ck_mental_models_subtype CHECK (subtype IN ('structural', 'emergent', 'pinned', 'learned'))
-        )
-    """)
-
-    # Step 5: Create indexes for efficient queries (if not exist)
-    op.execute(f"CREATE INDEX IF NOT EXISTS idx_mental_models_bank_id ON {schema}mental_models(bank_id)")
-    op.execute(f"CREATE INDEX IF NOT EXISTS idx_mental_models_subtype ON {schema}mental_models(bank_id, subtype)")
-    op.execute(f"CREATE INDEX IF NOT EXISTS idx_mental_models_entity_id ON {schema}mental_models(entity_id)")
-    # GIN index for efficient tags array filtering
-    op.execute(f"CREATE INDEX IF NOT EXISTS idx_mental_models_tags ON {schema}mental_models USING GIN(tags)")
-
-
-def downgrade() -> None:
-    """Revert mental models v4 changes."""
-    schema = _get_schema_prefix()
-
-    # Drop mental_models table (cascades to indexes)
-    op.execute(f"DROP TABLE IF EXISTS {schema}mental_models CASCADE")
-
-    # Add back background column to banks
-    op.execute(f"ALTER TABLE {schema}banks ADD COLUMN IF NOT EXISTS background TEXT")
-
-    # Migrate mission back to background
-    op.execute(f"UPDATE {schema}banks SET background = mission WHERE background IS NULL")
-
-    # Remove mission column
-    op.execute(f"ALTER TABLE {schema}banks DROP COLUMN IF EXISTS mission")
-
-    # Note: Cannot restore deleted observations - they are lost on downgrade
@@ -1,41 +0,0 @@
-"""delete_opinions
-
-Revision ID: i4d5e6f7g8h9
-Revises: h3c4d5e6f7g8
-Create Date: 2026-01-15 00:00:00.000000
-
-This migration removes opinion facts from memory_units.
-Opinions are no longer a separate fact type - they are now represented
-through mental model observations with confidence scores.
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-# revision identifiers, used by Alembic.
-revision: str = "i4d5e6f7g8h9"
-down_revision: str | Sequence[str] | None = "h3c4d5e6f7g8"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Delete opinion memory_units."""
-    schema = _get_schema_prefix()
-
-    # Delete opinion memory_units (cascades to unit_entities links)
-    # Opinions are now handled through mental model observations
-    op.execute(f"DELETE FROM {schema}memory_units WHERE fact_type = 'opinion'")
-
-
-def downgrade() -> None:
-    """Cannot restore deleted opinions."""
-    # Note: Cannot restore deleted opinions - they are lost on downgrade
-    pass
@@ -1,95 +0,0 @@
-"""mental_model_versions
-
-Revision ID: j5e6f7g8h9i0
-Revises: i4d5e6f7g8h9
-Create Date: 2026-01-16 00:00:00.000000
-
-This migration adds versioning support for mental models:
-1. Creates mental_model_versions table to store observation snapshots
-2. Adds version column to mental_models for tracking current version
-
-This enables changelog/diff functionality for mental model observations.
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-# revision identifiers, used by Alembic.
-revision: str = "j5e6f7g8h9i0"
-down_revision: str | Sequence[str] | None = "i4d5e6f7g8h9"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Create mental_model_versions table and add version tracking."""
-    schema = _get_schema_prefix()
-
-    # Create mental_model_versions table for storing observation snapshots
-    op.execute(f"""
-        CREATE TABLE {schema}mental_model_versions (
-            id SERIAL PRIMARY KEY,
-            mental_model_id VARCHAR(64) NOT NULL,
-            bank_id VARCHAR(64) NOT NULL,
-            version INT NOT NULL,
-            observations JSONB NOT NULL DEFAULT '{{"observations": []}}'::jsonb,
-            created_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT now(),
-            FOREIGN KEY (mental_model_id, bank_id)
-                REFERENCES {schema}mental_models(id, bank_id) ON DELETE CASCADE,
-            UNIQUE (mental_model_id, bank_id, version)
-        )
-    """)
-
-    # Index for efficient version queries (get latest, list versions)
-    op.execute(f"""
-        CREATE INDEX idx_mental_model_versions_lookup
-        ON {schema}mental_model_versions(mental_model_id, bank_id, version DESC)
-    """)
-
-    # Add version column to mental_models to track current version
-    op.execute(f"""
-        ALTER TABLE {schema}mental_models
-        ADD COLUMN IF NOT EXISTS version INT NOT NULL DEFAULT 0
-    """)
-
-    # Migrate existing mental models: create version 1 for any that have observations
-    op.execute(f"""
-        INSERT INTO {schema}mental_model_versions (mental_model_id, bank_id, version, observations, created_at)
-        SELECT id, bank_id, 1, observations, COALESCE(last_updated, created_at)
-        FROM {schema}mental_models
-        WHERE observations IS NOT NULL
-          AND observations != '{{"observations": []}}'::jsonb
-          AND (observations->'observations') IS NOT NULL
-          AND jsonb_array_length(observations->'observations') > 0
-    """)
-
-    # Update version to 1 for migrated mental models
-    op.execute(f"""
-        UPDATE {schema}mental_models
-        SET version = 1
-        WHERE observations IS NOT NULL
-          AND observations != '{{"observations": []}}'::jsonb
-          AND (observations->'observations') IS NOT NULL
-          AND jsonb_array_length(observations->'observations') > 0
-    """)
-
-
-def downgrade() -> None:
-    """Remove mental_model_versions table and version column."""
-    schema = _get_schema_prefix()
-
-    # Drop index
-    op.execute(f"DROP INDEX IF EXISTS {schema}idx_mental_model_versions_lookup")
-
-    # Drop versions table
-    op.execute(f"DROP TABLE IF EXISTS {schema}mental_model_versions")
-
-    # Remove version column from mental_models
-    op.execute(f"ALTER TABLE {schema}mental_models DROP COLUMN IF EXISTS version")
@@ -1,58 +0,0 @@
-"""add_directive_subtype
-
-Revision ID: k6f7g8h9i0j1
-Revises: j5e6f7g8h9i0
-Create Date: 2026-01-16 00:00:00.000000
-
-This migration adds 'directive' to the mental_models subtype constraint.
-Directives are hard rules with user-provided observations that the reflect agent must follow.
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-# revision identifiers, used by Alembic.
-revision: str = "k6f7g8h9i0j1"
-down_revision: str | Sequence[str] | None = "j5e6f7g8h9i0"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Add 'directive' to mental_models subtype constraint."""
-    schema = _get_schema_prefix()
-
-    # Drop existing constraint
-    op.execute(f"ALTER TABLE {schema}mental_models DROP CONSTRAINT IF EXISTS ck_mental_models_subtype")
-
-    # Create new constraint with 'directive' added
-    op.execute(f"""
-        ALTER TABLE {schema}mental_models
-        ADD CONSTRAINT ck_mental_models_subtype
-        CHECK (subtype IN ('structural', 'emergent', 'pinned', 'learned', 'directive'))
-    """)
-
-
-def downgrade() -> None:
-    """Remove 'directive' from mental_models subtype constraint."""
-    schema = _get_schema_prefix()
-
-    # First delete any directives (cannot downgrade if they exist)
-    op.execute(f"DELETE FROM {schema}mental_models WHERE subtype = 'directive'")
-
-    # Drop constraint with directive
-    op.execute(f"ALTER TABLE {schema}mental_models DROP CONSTRAINT IF EXISTS ck_mental_models_subtype")
-
-    # Recreate original constraint without directive
-    op.execute(f"""
-        ALTER TABLE {schema}mental_models
-        ADD CONSTRAINT ck_mental_models_subtype
-        CHECK (subtype IN ('structural', 'emergent', 'pinned', 'learned'))
-    """)
@@ -1,109 +0,0 @@
-"""add_worker_columns
-
-Revision ID: l7g8h9i0j1k2
-Revises: k6f7g8h9i0j1
-Create Date: 2026-01-19 00:00:00.000000
-
-This migration adds columns to async_operations for distributed worker support:
- worker_id: ID of the worker that claimed the task
- claimed_at: When the task was claimed
- retry_count: Number of retry attempts
- task_payload: The serialized task dictionary
-"""
-
-from collections.abc import Sequence
-
-import sqlalchemy as sa
-from alembic import context, op
-from sqlalchemy.dialects import postgresql
-
-# revision identifiers, used by Alembic.
-revision: str = "l7g8h9i0j1k2"
-down_revision: str | Sequence[str] | None = "k6f7g8h9i0j1"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Add worker columns to async_operations."""
-    schema = _get_schema_prefix()
-
-    # Add worker_id column (ID of worker that claimed the task)
-    op.add_column(
-        "async_operations",
-        sa.Column("worker_id", sa.Text(), nullable=True),
-        schema=context.config.get_main_option("target_schema") or None,
-    )
-
-    # Add claimed_at column (when task was claimed by worker)
-    op.add_column(
-        "async_operations",
-        sa.Column("claimed_at", postgresql.TIMESTAMP(timezone=True), nullable=True),
-        schema=context.config.get_main_option("target_schema") or None,
-    )
-
-    # Add retry_count column (number of retry attempts)
-    op.add_column(
-        "async_operations",
-        sa.Column("retry_count", sa.Integer(), server_default="0", nullable=False),
-        schema=context.config.get_main_option("target_schema") or None,
-    )
-
-    # Add task_payload column (serialized task dictionary)
-    op.add_column(
-        "async_operations",
-        sa.Column(
-            "task_payload",
-            postgresql.JSONB(astext_type=sa.Text()),
-            nullable=True,
-        ),
-        schema=context.config.get_main_option("target_schema") or None,
-    )
-
-    # Add index for efficient worker polling (pending tasks ordered by creation time)
-    op.execute(
-        f"CREATE INDEX idx_async_operations_pending_claim ON {schema}async_operations (status, created_at) "
-        f"WHERE status = 'pending' AND task_payload IS NOT NULL"
-    )
-
-    # Add index for finding tasks by worker_id (for decommissioning)
-    op.execute(
-        f"CREATE INDEX idx_async_operations_worker_id ON {schema}async_operations (worker_id) WHERE worker_id IS NOT NULL"
-    )
-
-
-def downgrade() -> None:
-    """Remove worker columns from async_operations."""
-    schema = _get_schema_prefix()
-
-    # Drop indexes
-    op.execute(f"DROP INDEX IF EXISTS {schema}idx_async_operations_pending_claim")
-    op.execute(f"DROP INDEX IF EXISTS {schema}idx_async_operations_worker_id")
-
-    # Drop columns
-    op.drop_column(
-        "async_operations",
-        "task_payload",
-        schema=context.config.get_main_option("target_schema") or None,
-    )
-    op.drop_column(
-        "async_operations",
-        "retry_count",
-        schema=context.config.get_main_option("target_schema") or None,
-    )
-    op.drop_column(
-        "async_operations",
-        "claimed_at",
-        schema=context.config.get_main_option("target_schema") or None,
-    )
-    op.drop_column(
-        "async_operations",
-        "worker_id",
-        schema=context.config.get_main_option("target_schema") or None,
-    )
@@ -1,41 +0,0 @@
-"""mental_model_id_to_text
-
-Revision ID: m8h9i0j1k2l3
-Revises: l7g8h9i0j1k2
-Create Date: 2026-01-19 00:00:00.000000
-
-This migration changes the mental_models.id column from VARCHAR(64) to TEXT
-to support longer model IDs (e.g., entity names that exceed 64 characters).
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-# revision identifiers, used by Alembic.
-revision: str = "m8h9i0j1k2l3"
-down_revision: str | Sequence[str] | None = "l7g8h9i0j1k2"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Change mental_models.id from VARCHAR(64) to TEXT."""
-    schema = _get_schema_prefix()
-
-    # Alter the id column type from VARCHAR(64) to TEXT
-    op.execute(f"ALTER TABLE {schema}mental_models ALTER COLUMN id TYPE TEXT")
-
-
-def downgrade() -> None:
-    """Revert mental_models.id from TEXT to VARCHAR(64)."""
-    schema = _get_schema_prefix()
-
-    # Note: This may fail if any id values exceed 64 characters
-    op.execute(f"ALTER TABLE {schema}mental_models ALTER COLUMN id TYPE VARCHAR(64)")
@@ -1,134 +0,0 @@
-"""learnings_and_pinned_reflections
-
-Revision ID: n9i0j1k2l3m4
-Revises: m8h9i0j1k2l3
-Create Date: 2026-01-21 00:00:00.000000
-
-This migration:
-1. Creates the 'learnings' table for automatic bottom-up consolidation
-2. Creates the 'pinned_reflections' table for user-curated living documents
-3. Adds consolidation tracking columns to the 'banks' table
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-# revision identifiers, used by Alembic.
-revision: str = "n9i0j1k2l3m4"
-down_revision: str | Sequence[str] | None = "m8h9i0j1k2l3"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Create learnings and pinned_reflections tables."""
-    schema = _get_schema_prefix()
-
-    # 1. Create learnings table
-    op.execute(f"""
-        CREATE TABLE {schema}learnings (
-            id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
-            bank_id VARCHAR(64) NOT NULL,
-            text TEXT NOT NULL,
-            proof_count INT NOT NULL DEFAULT 1,
-            history JSONB DEFAULT '[]'::jsonb,
-            mission_context VARCHAR(64),
-            pre_mission_change BOOLEAN DEFAULT FALSE,
-            embedding vector(384),
-            tags VARCHAR[] DEFAULT ARRAY[]::VARCHAR[],
-            created_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT now(),
-            updated_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT now()
-        )
-    """)
-
-    # Add foreign key constraint
-    op.execute(f"""
-        ALTER TABLE {schema}learnings
-        ADD CONSTRAINT fk_learnings_bank_id
-        FOREIGN KEY (bank_id) REFERENCES {schema}banks(bank_id) ON DELETE CASCADE
-    """)
-
-    # Indexes for learnings
-    op.execute(f"CREATE INDEX idx_learnings_bank_id ON {schema}learnings(bank_id)")
-    op.execute(f"""
-        CREATE INDEX idx_learnings_embedding ON {schema}learnings
-        USING hnsw (embedding vector_cosine_ops)
-    """)
-    op.execute(f"CREATE INDEX idx_learnings_tags ON {schema}learnings USING GIN(tags)")
-
-    # Full-text search for learnings
-    op.execute(f"""
-        ALTER TABLE {schema}learnings ADD COLUMN search_vector tsvector
-        GENERATED ALWAYS AS (to_tsvector('english', text)) STORED
-    """)
-    op.execute(f"CREATE INDEX idx_learnings_text_search ON {schema}learnings USING gin(search_vector)")
-
-    # 2. Create pinned_reflections table
-    op.execute(f"""
-        CREATE TABLE {schema}pinned_reflections (
-            id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
-            bank_id VARCHAR(64) NOT NULL,
-            name VARCHAR(256) NOT NULL,
-            source_query TEXT NOT NULL,
-            content TEXT NOT NULL,
-            embedding vector(384),
-            tags VARCHAR[] DEFAULT ARRAY[]::VARCHAR[],
-            last_refreshed_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT now(),
-            created_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT now()
-        )
-    """)
-
-    # Add foreign key constraint
-    op.execute(f"""
-        ALTER TABLE {schema}pinned_reflections
-        ADD CONSTRAINT fk_pinned_reflections_bank_id
-        FOREIGN KEY (bank_id) REFERENCES {schema}banks(bank_id) ON DELETE CASCADE
-    """)
-
-    # Indexes for pinned_reflections
-    op.execute(f"CREATE INDEX idx_pinned_reflections_bank_id ON {schema}pinned_reflections(bank_id)")
-    op.execute(f"""
-        CREATE INDEX idx_pinned_reflections_embedding ON {schema}pinned_reflections
-        USING hnsw (embedding vector_cosine_ops)
-    """)
-    op.execute(f"CREATE INDEX idx_pinned_reflections_tags ON {schema}pinned_reflections USING GIN(tags)")
-
-    # Full-text search for pinned_reflections
-    op.execute(f"""
-        ALTER TABLE {schema}pinned_reflections ADD COLUMN search_vector tsvector
-        GENERATED ALWAYS AS (to_tsvector('english', COALESCE(name, '') || ' ' || content)) STORED
-    """)
-    op.execute(f"""
-        CREATE INDEX idx_pinned_reflections_text_search ON {schema}pinned_reflections
-        USING gin(search_vector)
-    """)
-
-    # 3. Add consolidation tracking columns to banks table
-    op.execute(f"""
-        ALTER TABLE {schema}banks
-        ADD COLUMN IF NOT EXISTS last_consolidated_at TIMESTAMP WITH TIME ZONE
-    """)
-    op.execute(f"""
-        ALTER TABLE {schema}banks
-        ADD COLUMN IF NOT EXISTS mission_changed_at TIMESTAMP WITH TIME ZONE
-    """)
-
-
-def downgrade() -> None:
-    """Drop learnings and pinned_reflections tables."""
-    schema = _get_schema_prefix()
-
-    # Drop tables
-    op.execute(f"DROP TABLE IF EXISTS {schema}learnings CASCADE")
-    op.execute(f"DROP TABLE IF EXISTS {schema}pinned_reflections CASCADE")
-
-    # Remove columns from banks
-    op.execute(f"ALTER TABLE {schema}banks DROP COLUMN IF EXISTS last_consolidated_at")
-    op.execute(f"ALTER TABLE {schema}banks DROP COLUMN IF EXISTS mission_changed_at")
@@ -1,113 +0,0 @@
-"""migrate_mental_models_data
-
-Revision ID: o0j1k2l3m4n5
-Revises: n9i0j1k2l3m4
-Create Date: 2026-01-21 00:00:00.000000
-
-This migration:
-1. Migrates existing 'pinned' mental models to the new 'pinned_reflections' table
-2. Migrates existing 'learned' mental models to the new 'learnings' table
-3. Deletes non-directive mental models (structural, emergent, pinned, learned)
-4. Drops the mental_model_versions table (no longer used)
-5. Adds a CHECK constraint that only 'directive' subtype is allowed
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-# revision identifiers, used by Alembic.
-revision: str = "o0j1k2l3m4n5"
-down_revision: str | Sequence[str] | None = "n9i0j1k2l3m4"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Migrate data and clean up old mental models."""
-    schema = _get_schema_prefix()
-
-    # 1. Migrate 'pinned' mental models to pinned_reflections
-    # For pinned models, the first observation's content becomes the pinned reflection content
-    op.execute(f"""
-        INSERT INTO {schema}pinned_reflections (bank_id, name, source_query, content, tags, created_at)
-        SELECT
-            bank_id,
-            name,
-            description AS source_query,
-            COALESCE(
-                observations->'observations'->0->>'content',
-                description,
-                ''
-            ) AS content,
-            tags,
-            created_at
-        FROM {schema}mental_models
-        WHERE subtype = 'pinned'
-        ON CONFLICT DO NOTHING
-    """)
-
-    # 2. Migrate 'learned' mental models to learnings
-    # Each observation in a learned model becomes a separate learning
-    op.execute(f"""
-        INSERT INTO {schema}learnings (bank_id, text, proof_count, tags, created_at)
-        SELECT
-            mm.bank_id,
-            obs->>'content' AS text,
-            GREATEST(1, COALESCE(jsonb_array_length(obs->'evidence'), 1)) AS proof_count,
-            mm.tags,
-            mm.created_at
-        FROM {schema}mental_models mm,
-        LATERAL jsonb_array_elements(mm.observations->'observations') AS obs
-        WHERE mm.subtype = 'learned'
-          AND obs->>'content' IS NOT NULL
-          AND obs->>'content' != ''
-        ON CONFLICT DO NOTHING
-    """)
-
-    # 3. Delete all non-directive mental models (they've been migrated or are obsolete)
-    op.execute(f"""
-        DELETE FROM {schema}mental_models
-        WHERE subtype != 'directive'
-    """)
-
-    # 4. Drop the mental_model_versions table (no longer used)
-    op.execute(f"DROP TABLE IF EXISTS {schema}mental_model_versions CASCADE")
-
-    # 5. Drop old constraints and add new one that only allows 'directive'
-    op.execute(f"ALTER TABLE {schema}mental_models DROP CONSTRAINT IF EXISTS ck_mental_models_subtype")
-    op.execute(f"""
-        ALTER TABLE {schema}mental_models
-        ADD CONSTRAINT ck_mental_models_subtype CHECK (subtype = 'directive')
-    """)
-
-
-def downgrade() -> None:
-    """Reverse the migration (data migration is one-way, so this just removes constraints)."""
-    schema = _get_schema_prefix()
-
-    # Remove the directive-only constraint
-    op.execute(f"ALTER TABLE {schema}mental_models DROP CONSTRAINT IF EXISTS ck_mental_models_subtype")
-
-    # Re-create mental_model_versions table
-    op.execute(f"""
-        CREATE TABLE IF NOT EXISTS {schema}mental_model_versions (
-            id SERIAL PRIMARY KEY,
-            bank_id VARCHAR(64) NOT NULL,
-            model_id VARCHAR(128) NOT NULL,
-            version INT NOT NULL,
-            observations JSONB NOT NULL,
-            created_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT now()
-        )
-    """)
-    op.execute(
-        f"CREATE INDEX IF NOT EXISTS idx_mm_versions_lookup ON {schema}mental_model_versions(bank_id, model_id, version DESC)"
-    )
-
-    # Note: Data migration cannot be reversed - pinned_reflections and learnings data remains
@@ -1,194 +0,0 @@
-"""new_knowledge_architecture
-
-Revision ID: p1k2l3m4n5o6
-Revises: o0j1k2l3m4n5
-Create Date: 2026-01-21 00:00:00.000000
-
-This migration implements the new knowledge architecture:
-1. Drops the 'learnings' table (mental models are now in memory_units)
-2. Renames 'pinned_reflections' to 'reflections'
-3. Drops the 'mental_models' table completely
-4. Creates 'directives' table for hard rules
-5. Adds mental model support columns to 'memory_units' (proof_count, source_memory_ids, history)
-
-The new architecture:
- Directives: Hard rules in their own table
- Mental Models: Stored in memory_units with fact_type='mental_model'
- Reflections: User-curated documents (renamed from pinned_reflections)
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-# revision identifiers, used by Alembic.
-revision: str = "p1k2l3m4n5o6"
-down_revision: str | Sequence[str] | None = "o0j1k2l3m4n5"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Implement new knowledge architecture."""
-    schema = _get_schema_prefix()
-
-    # 1. Drop the learnings table (mental models will be in memory_units)
-    op.execute(f"DROP TABLE IF EXISTS {schema}learnings CASCADE")
-
-    # 2. Rename pinned_reflections to reflections
-    op.execute(f"ALTER TABLE IF EXISTS {schema}pinned_reflections RENAME TO reflections")
-
-    # Rename indexes for reflections
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_pinned_reflections_bank_id RENAME TO idx_reflections_bank_id")
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_pinned_reflections_embedding RENAME TO idx_reflections_embedding")
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_pinned_reflections_tags RENAME TO idx_reflections_tags")
-    op.execute(
-        f"ALTER INDEX IF EXISTS {schema}idx_pinned_reflections_text_search RENAME TO idx_reflections_text_search"
-    )
-
-    # Rename foreign key constraint
-    op.execute(f"""
-        ALTER TABLE {schema}reflections
-        DROP CONSTRAINT IF EXISTS fk_pinned_reflections_bank_id
-    """)
-    op.execute(f"""
-        ALTER TABLE {schema}reflections
-        ADD CONSTRAINT fk_reflections_bank_id
-        FOREIGN KEY (bank_id) REFERENCES {schema}banks(bank_id) ON DELETE CASCADE
-    """)
-
-    # 3. Drop the mental_models table completely
-    op.execute(f"DROP TABLE IF EXISTS {schema}mental_models CASCADE")
-
-    # 4. Create directives table
-    op.execute(f"""
-        CREATE TABLE {schema}directives (
-            id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
-            bank_id VARCHAR(64) NOT NULL,
-            name VARCHAR(256) NOT NULL,
-            content TEXT NOT NULL,
-            priority INT NOT NULL DEFAULT 0,
-            is_active BOOLEAN NOT NULL DEFAULT TRUE,
-            tags VARCHAR[] DEFAULT ARRAY[]::VARCHAR[],
-            created_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT now(),
-            updated_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT now()
-        )
-    """)
-
-    # Add foreign key and indexes for directives
-    op.execute(f"""
-        ALTER TABLE {schema}directives
-        ADD CONSTRAINT fk_directives_bank_id
-        FOREIGN KEY (bank_id) REFERENCES {schema}banks(bank_id) ON DELETE CASCADE
-    """)
-    op.execute(f"CREATE INDEX idx_directives_bank_id ON {schema}directives(bank_id)")
-    op.execute(f"CREATE INDEX idx_directives_bank_active ON {schema}directives(bank_id, is_active)")
-    op.execute(f"CREATE INDEX idx_directives_tags ON {schema}directives USING GIN(tags)")
-
-    # 5. Add mental model support columns to memory_units
-    # proof_count: Number of memories that support this mental model
-    op.execute(f"""
-        ALTER TABLE {schema}memory_units
-        ADD COLUMN IF NOT EXISTS proof_count INT DEFAULT 1
-    """)
-
-    # source_memory_ids: Array of memory IDs that consolidated into this mental model
-    op.execute(f"""
-        ALTER TABLE {schema}memory_units
-        ADD COLUMN IF NOT EXISTS source_memory_ids UUID[] DEFAULT ARRAY[]::UUID[]
-    """)
-
-    # history: JSONB array tracking changes to mental models
-    op.execute(f"""
-        ALTER TABLE {schema}memory_units
-        ADD COLUMN IF NOT EXISTS history JSONB DEFAULT '[]'::jsonb
-    """)
-
-    # Add index for finding mental models
-    op.execute(f"""
-        CREATE INDEX IF NOT EXISTS idx_memory_units_mental_models
-        ON {schema}memory_units(bank_id, fact_type)
-        WHERE fact_type = 'mental_model'
-    """)
-
-    # 6. Update fact_type check constraint to include 'mental_model'
-    op.execute(f"ALTER TABLE {schema}memory_units DROP CONSTRAINT IF EXISTS memory_units_fact_type_check")
-    op.execute(f"""
-        ALTER TABLE {schema}memory_units
-        ADD CONSTRAINT memory_units_fact_type_check
-        CHECK (fact_type IN ('world', 'experience', 'opinion', 'observation', 'mental_model'))
-    """)
-
-
-def downgrade() -> None:
-    """Reverse the migration."""
-    schema = _get_schema_prefix()
-
-    # Restore original fact_type check constraint (without 'mental_model')
-    op.execute(f"ALTER TABLE {schema}memory_units DROP CONSTRAINT IF EXISTS memory_units_fact_type_check")
-    op.execute(f"""
-        ALTER TABLE {schema}memory_units
-        ADD CONSTRAINT memory_units_fact_type_check
-        CHECK (fact_type IN ('world', 'experience', 'opinion', 'observation'))
-    """)
-
-    # Drop mental model columns from memory_units
-    op.execute(f"ALTER TABLE {schema}memory_units DROP COLUMN IF EXISTS proof_count")
-    op.execute(f"ALTER TABLE {schema}memory_units DROP COLUMN IF EXISTS source_memory_ids")
-    op.execute(f"ALTER TABLE {schema}memory_units DROP COLUMN IF EXISTS history")
-    op.execute(f"DROP INDEX IF EXISTS {schema}idx_memory_units_mental_models")
-
-    # Drop directives table
-    op.execute(f"DROP TABLE IF EXISTS {schema}directives CASCADE")
-
-    # Rename reflections back to pinned_reflections
-    op.execute(f"ALTER TABLE IF EXISTS {schema}reflections RENAME TO pinned_reflections")
-
-    # Restore indexes
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_reflections_bank_id RENAME TO idx_pinned_reflections_bank_id")
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_reflections_embedding RENAME TO idx_pinned_reflections_embedding")
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_reflections_tags RENAME TO idx_pinned_reflections_tags")
-    op.execute(
-        f"ALTER INDEX IF EXISTS {schema}idx_reflections_text_search RENAME TO idx_pinned_reflections_text_search"
-    )
-
-    # Restore foreign key
-    op.execute(f"""
-        ALTER TABLE {schema}pinned_reflections
-        DROP CONSTRAINT IF EXISTS fk_reflections_bank_id
-    """)
-    op.execute(f"""
-        ALTER TABLE {schema}pinned_reflections
-        ADD CONSTRAINT fk_pinned_reflections_bank_id
-        FOREIGN KEY (bank_id) REFERENCES {schema}banks(bank_id) ON DELETE CASCADE
-    """)
-
-    # Re-create learnings table
-    op.execute(f"""
-        CREATE TABLE IF NOT EXISTS {schema}learnings (
-            id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
-            bank_id VARCHAR(64) NOT NULL,
-            text TEXT NOT NULL,
-            proof_count INT NOT NULL DEFAULT 1,
-            history JSONB DEFAULT '[]'::jsonb,
-            mission_context VARCHAR(64),
-            pre_mission_change BOOLEAN DEFAULT FALSE,
-            embedding vector(384),
-            tags VARCHAR[] DEFAULT ARRAY[]::VARCHAR[],
-            created_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT now(),
-            updated_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT now()
-        )
-    """)
-    op.execute(f"""
-        ALTER TABLE {schema}learnings
-        ADD CONSTRAINT fk_learnings_bank_id
-        FOREIGN KEY (bank_id) REFERENCES {schema}banks(bank_id) ON DELETE CASCADE
-    """)
-
-    # Note: mental_models table recreation is complex and would need separate handling
@@ -1,50 +0,0 @@
-"""fix_mental_model_fact_type
-
-Revision ID: q2l3m4n5o6p7
-Revises: p1k2l3m4n5o6
-Create Date: 2026-01-21 13:30:00.000000
-
-Fix the fact_type check constraint to include 'mental_model'.
-This is a fix for p1k2l3m4n5o6 which should have included this change.
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-# revision identifiers, used by Alembic.
-revision: str = "q2l3m4n5o6p7"
-down_revision: str | Sequence[str] | None = "p1k2l3m4n5o6"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Add 'mental_model' to the fact_type check constraint."""
-    schema = _get_schema_prefix()
-
-    # Drop the old constraint and add the new one with mental_model included
-    op.execute(f"ALTER TABLE {schema}memory_units DROP CONSTRAINT IF EXISTS memory_units_fact_type_check")
-    op.execute(f"""
-        ALTER TABLE {schema}memory_units
-        ADD CONSTRAINT memory_units_fact_type_check
-        CHECK (fact_type IN ('world', 'experience', 'opinion', 'observation', 'mental_model'))
-    """)
-
-
-def downgrade() -> None:
-    """Remove 'mental_model' from the fact_type check constraint."""
-    schema = _get_schema_prefix()
-
-    op.execute(f"ALTER TABLE {schema}memory_units DROP CONSTRAINT IF EXISTS memory_units_fact_type_check")
-    op.execute(f"""
-        ALTER TABLE {schema}memory_units
-        ADD CONSTRAINT memory_units_fact_type_check
-        CHECK (fact_type IN ('world', 'experience', 'opinion', 'observation'))
-    """)
@@ -1,47 +0,0 @@
-"""Add reflect_response JSONB column to reflections
-
-Revision ID: r3m4n5o6p7q8
-Revises: q2l3m4n5o6p7
-Create Date: 2026-01-21
-
-This migration adds a reflect_response JSONB column to store the full
-reflect API response payload, including based_on facts and trace data.
-
-Note: Table was renamed from pinned_reflections to reflections in p1k2l3m4n5o6.
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-revision: str = "r3m4n5o6p7q8"
-down_revision: str | Sequence[str] | None = "q2l3m4n5o6p7"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Add reflect_response JSONB column to reflections."""
-    schema = _get_schema_prefix()
-
-    # Add reflect_response column to store the full reflect API response
-    op.execute(f"""
-        ALTER TABLE {schema}reflections
-        ADD COLUMN IF NOT EXISTS reflect_response JSONB
-    """)
-
-
-def downgrade() -> None:
-    """Remove reflect_response column from reflections."""
-    schema = _get_schema_prefix()
-
-    op.execute(f"""
-        ALTER TABLE {schema}reflections
-        DROP COLUMN IF EXISTS reflect_response
-    """)
@@ -1,53 +0,0 @@
-"""Add consolidated_at column to memory_units for incremental consolidation tracking.
-
-This allows consolidation to track progress at the memory level rather than
-using a bank-level watermark. If consolidation crashes, already-processed
-memories won't be reprocessed.
-
-Revision ID: s4n5o6p7q8r9
-Revises: r3m4n5o6p7q8
-Create Date: 2025-01-22
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-revision: str = "s4n5o6p7q8r9"
-down_revision: str | Sequence[str] | None = "r3m4n5o6p7q8"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    schema = _get_schema_prefix()
-
-    # Add consolidated_at column to memory_units
-    op.execute(
-        f"""
-        ALTER TABLE {schema}memory_units
-        ADD COLUMN IF NOT EXISTS consolidated_at TIMESTAMPTZ DEFAULT NULL
-        """
-    )
-
-    # Create index for efficient querying of unconsolidated memories
-    op.execute(
-        f"""
-        CREATE INDEX IF NOT EXISTS idx_memory_units_unconsolidated
-        ON {schema}memory_units (bank_id, created_at)
-        WHERE consolidated_at IS NULL AND fact_type IN ('experience', 'world')
-        """
-    )
-
-
-def downgrade() -> None:
-    schema = _get_schema_prefix()
-
-    op.execute(f"DROP INDEX IF EXISTS {schema}idx_memory_units_unconsolidated")
-    op.execute(f"ALTER TABLE {schema}memory_units DROP COLUMN IF EXISTS consolidated_at")
@@ -1,134 +0,0 @@
-"""Rename mental_model fact_type to observation and reflections table to mental_models
-
-Revision ID: t5o6p7q8r9s0
-Revises: s4n5o6p7q8r9
-Create Date: 2026-01-26
-
-This migration implements the terminology rename:
-1. mental_model (fact_type in memory_units) -> observation
-2. reflections table -> mental_models table
-
-The new terminology:
- Observations: Consolidated knowledge synthesized from facts (was mental_model)
- Mental Models: Stored reflect responses (was reflections)
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-revision: str = "t5o6p7q8r9s0"
-down_revision: str | Sequence[str] | None = "s4n5o6p7q8r9"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Rename mental_model -> observation and reflections -> mental_models."""
-    schema = _get_schema_prefix()
-
-    # 1. Update fact_type values: mental_model -> observation
-    op.execute(f"""
-        UPDATE {schema}memory_units
-        SET fact_type = 'observation'
-        WHERE fact_type = 'mental_model'
-    """)
-
-    # 2. Update the CHECK constraint - remove mental_model, keep observation
-    op.execute(f"ALTER TABLE {schema}memory_units DROP CONSTRAINT IF EXISTS memory_units_fact_type_check")
-    op.execute(f"""
-        ALTER TABLE {schema}memory_units
-        ADD CONSTRAINT memory_units_fact_type_check
-        CHECK (fact_type IN ('world', 'experience', 'opinion', 'observation'))
-    """)
-
-    # 3. Rename the index for observations (was for mental_models)
-    op.execute(f"DROP INDEX IF EXISTS {schema}idx_memory_units_mental_models")
-    op.execute(f"""
-        CREATE INDEX IF NOT EXISTS idx_memory_units_observations
-        ON {schema}memory_units(bank_id, fact_type)
-        WHERE fact_type = 'observation'
-    """)
-
-    # 4. Update the unconsolidated index to not filter by fact_type since observations
-    # are now the consolidated type
-    op.execute(f"DROP INDEX IF EXISTS {schema}idx_memory_units_unconsolidated")
-    op.execute(f"""
-        CREATE INDEX IF NOT EXISTS idx_memory_units_unconsolidated
-        ON {schema}memory_units (bank_id, created_at)
-        WHERE consolidated_at IS NULL AND fact_type IN ('experience', 'world')
-    """)
-
-    # 5. Rename reflections table to mental_models
-    op.execute(f"ALTER TABLE IF EXISTS {schema}reflections RENAME TO mental_models")
-
-    # 6. Rename indexes for mental_models (was reflections)
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_reflections_bank_id RENAME TO idx_mental_models_bank_id")
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_reflections_embedding RENAME TO idx_mental_models_embedding")
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_reflections_tags RENAME TO idx_mental_models_tags")
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_reflections_text_search RENAME TO idx_mental_models_text_search")
-
-    # 7. Rename foreign key constraint
-    op.execute(f"""
-        ALTER TABLE {schema}mental_models
-        DROP CONSTRAINT IF EXISTS fk_reflections_bank_id
-    """)
-    op.execute(f"""
-        ALTER TABLE {schema}mental_models
-        ADD CONSTRAINT fk_mental_models_bank_id
-        FOREIGN KEY (bank_id) REFERENCES {schema}banks(bank_id) ON DELETE CASCADE
-    """)
-
-
-def downgrade() -> None:
-    """Reverse: observation -> mental_model and mental_models -> reflections."""
-    schema = _get_schema_prefix()
-
-    # 1. Rename mental_models table back to reflections
-    op.execute(f"ALTER TABLE IF EXISTS {schema}mental_models RENAME TO reflections")
-
-    # 2. Rename indexes back
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_mental_models_bank_id RENAME TO idx_reflections_bank_id")
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_mental_models_embedding RENAME TO idx_reflections_embedding")
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_mental_models_tags RENAME TO idx_reflections_tags")
-    op.execute(f"ALTER INDEX IF EXISTS {schema}idx_mental_models_text_search RENAME TO idx_reflections_text_search")
-
-    # 3. Rename foreign key back
-    op.execute(f"""
-        ALTER TABLE {schema}reflections
-        DROP CONSTRAINT IF EXISTS fk_mental_models_bank_id
-    """)
-    op.execute(f"""
-        ALTER TABLE {schema}reflections
-        ADD CONSTRAINT fk_reflections_bank_id
-        FOREIGN KEY (bank_id) REFERENCES {schema}banks(bank_id) ON DELETE CASCADE
-    """)
-
-    # 4. Update fact_type values: observation -> mental_model
-    op.execute(f"""
-        UPDATE {schema}memory_units
-        SET fact_type = 'mental_model'
-        WHERE fact_type = 'observation'
-    """)
-
-    # 5. Update the CHECK constraint back
-    op.execute(f"ALTER TABLE {schema}memory_units DROP CONSTRAINT IF EXISTS memory_units_fact_type_check")
-    op.execute(f"""
-        ALTER TABLE {schema}memory_units
-        ADD CONSTRAINT memory_units_fact_type_check
-        CHECK (fact_type IN ('world', 'experience', 'opinion', 'observation', 'mental_model'))
-    """)
-
-    # 6. Rename index back
-    op.execute(f"DROP INDEX IF EXISTS {schema}idx_memory_units_observations")
-    op.execute(f"""
-        CREATE INDEX IF NOT EXISTS idx_memory_units_mental_models
-        ON {schema}memory_units(bank_id, fact_type)
-        WHERE fact_type = 'mental_model'
-    """)
@@ -1,41 +0,0 @@
-"""Change mental_models.id from UUID to TEXT
-
-Revision ID: u6p7q8r9s0t1
-Revises: t5o6p7q8r9s0
-Create Date: 2026-01-27
-
-This migration changes the mental_models.id column from UUID to TEXT
-to support user-defined text identifiers like 'team-communication' instead of UUIDs.
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-revision: str = "u6p7q8r9s0t1"
-down_revision: str | Sequence[str] | None = "t5o6p7q8r9s0"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Change mental_models.id from UUID to TEXT."""
-    schema = _get_schema_prefix()
-
-    # Change the id column type from UUID to TEXT
-    # Existing UUIDs will be converted to their string representation
-    op.execute(f"ALTER TABLE {schema}mental_models ALTER COLUMN id TYPE TEXT USING id::TEXT")
-
-
-def downgrade() -> None:
-    """Revert mental_models.id from TEXT to UUID."""
-    schema = _get_schema_prefix()
-
-    # Note: This will fail if any id values are not valid UUIDs
-    op.execute(f"ALTER TABLE {schema}mental_models ALTER COLUMN id TYPE UUID USING id::UUID")
@@ -1,50 +0,0 @@
-"""Add max_tokens and trigger columns to mental_models
-
-Revision ID: v7q8r9s0t1u2
-Revises: u6p7q8r9s0t1
-Create Date: 2026-01-27
-
-This migration adds:
- max_tokens column: token limit for content generation during refresh
- trigger column: JSONB for trigger settings (e.g., refresh_after_consolidation)
-"""
-
-from collections.abc import Sequence
-
-from alembic import context, op
-
-revision: str = "v7q8r9s0t1u2"
-down_revision: str | Sequence[str] | None = "u6p7q8r9s0t1"
-branch_labels: str | Sequence[str] | None = None
-depends_on: str | Sequence[str] | None = None
-
-
-def _get_schema_prefix() -> str:
-    """Get schema prefix for table names (required for multi-tenant support)."""
-    schema = context.config.get_main_option("target_schema")
-    return f'"{schema}".' if schema else ""
-
-
-def upgrade() -> None:
-    """Add max_tokens and trigger columns to mental_models."""
-    schema = _get_schema_prefix()
-
-    op.execute(f"""
-        ALTER TABLE {schema}mental_models
-        ADD COLUMN IF NOT EXISTS max_tokens INT NOT NULL DEFAULT 2048
-    """)
-
-    # trigger column stores trigger settings as JSONB
-    # Default: refresh_after_consolidation = false (not "real time")
-    op.execute(f"""
-        ALTER TABLE {schema}mental_models
-        ADD COLUMN IF NOT EXISTS trigger JSONB NOT NULL DEFAULT '{{"refresh_after_consolidation": false}}'::jsonb
-    """)
-
-
-def downgrade() -> None:
-    """Remove max_tokens and trigger columns from mental_models."""
-    schema = _get_schema_prefix()
-
-    op.execute(f"ALTER TABLE {schema}mental_models DROP COLUMN IF EXISTS max_tokens")
-    op.execute(f"ALTER TABLE {schema}mental_models DROP COLUMN IF EXISTS trigger")
@@ -1,4 +1,4 @@
-"""Hindsight MCP Server implementation using FastMCP (HTTP transport)."""
+"""Hindsight MCP Server implementation using FastMCP."""

 import json
 import logging
@@ -8,7 +8,8 @@ from contextvars import ContextVar
 from fastmcp import FastMCP

 from hindsight_api import MemoryEngine
-from hindsight_api.mcp_tools import MCPToolsConfig, register_mcp_tools
+from hindsight_api.engine.response_models import VALID_RECALL_FACT_TYPES
+from hindsight_api.models import RequestContext

 # Configure logging from HINDSIGHT_API_LOG_LEVEL environment variable
 _log_level_str = os.environ.get("HINDSIGHT_API_LOG_LEVEL", "info").lower()
@@ -51,15 +52,194 @@ def create_mcp_server(memory: MemoryEngine) -> FastMCP:
    # Use stateless_http=True for Claude Code compatibility
    mcp = FastMCP("hindsight-mcp-server", stateless_http=True)

-    # Configure and register tools using shared module
-    config = MCPToolsConfig(
-        bank_id_resolver=get_current_bank_id,
-        include_bank_id_param=True,  # HTTP MCP supports multi-bank via parameter
-        tools=None,  # All tools
-        retain_fire_and_forget=False,  # HTTP MCP supports sync/async modes
-    )
+    @mcp.tool()
+    async def retain(
+        content: str,
+        context: str = "general",
+        async_processing: bool = True,
+        bank_id: str | None = None,
+    ) -> str:
+        """
+        Store important information to long-term memory.

-    register_mcp_tools(mcp, memory, config)
+        Use this tool PROACTIVELY whenever the user shares:
+        - Personal facts, preferences, or interests
+        - Important events or milestones
+        - User history, experiences, or background
+        - Decisions, opinions, or stated preferences
+        - Goals, plans, or future intentions
+        - Relationships or people mentioned
+        - Work context, projects, or responsibilities
+
+        Args:
+            content: The fact/memory to store (be specific and include relevant details)
+            context: Category for the memory (e.g., 'preferences', 'work', 'hobbies', 'family'). Default: 'general'
+            async_processing: If True, queue for background processing and return immediately. If False, wait for completion. Default: True
+            bank_id: Optional bank to store in (defaults to session bank). Use for cross-bank operations.
+        """
+        try:
+            target_bank = bank_id or get_current_bank_id()
+            if target_bank is None:
+                return "Error: No bank_id configured"
+            contents = [{"content": content, "context": context}]
+            if async_processing:
+                # Queue for background processing and return immediately
+                result = await memory.submit_async_retain(
+                    bank_id=target_bank, contents=contents, request_context=RequestContext()
+                )
+                return f"Memory queued for background processing (operation_id: {result.get('operation_id', 'N/A')})"
+            else:
+                # Wait for completion
+                await memory.retain_batch_async(
+                    bank_id=target_bank,
+                    contents=contents,
+                    request_context=RequestContext(),
+                )
+                return f"Memory stored successfully in bank '{target_bank}'"
+        except Exception as e:
+            logger.error(f"Error storing memory: {e}", exc_info=True)
+            return f"Error: {str(e)}"
+
+    @mcp.tool()
+    async def recall(query: str, max_tokens: int = 4096, bank_id: str | None = None) -> str:
+        """
+        Search memories to provide personalized, context-aware responses.
+
+        Use this tool PROACTIVELY to:
+        - Check user's preferences before making suggestions
+        - Recall user's history to provide continuity
+        - Remember user's goals and context
+        - Personalize responses based on past interactions
+
+        Args:
+            query: Natural language search query (e.g., "user's food preferences", "what projects is user working on")
+            max_tokens: Maximum tokens in the response (default: 4096)
+            bank_id: Optional bank to search in (defaults to session bank). Use for cross-bank operations.
+        """
+        try:
+            target_bank = bank_id or get_current_bank_id()
+            if target_bank is None:
+                return "Error: No bank_id configured"
+            from hindsight_api.engine.memory_engine import Budget
+
+            recall_result = await memory.recall_async(
+                bank_id=target_bank,
+                query=query,
+                fact_type=list(VALID_RECALL_FACT_TYPES),
+                budget=Budget.HIGH,
+                max_tokens=max_tokens,
+                request_context=RequestContext(),
+            )
+
+            # Use model's JSON serialization
+            return recall_result.model_dump_json(indent=2)
+        except Exception as e:
+            logger.error(f"Error searching: {e}", exc_info=True)
+            return f'{{"error": "{e}", "results": []}}'
+
+    @mcp.tool()
+    async def reflect(query: str, context: str | None = None, budget: str = "low", bank_id: str | None = None) -> str:
+        """
+        Generate thoughtful analysis by synthesizing stored memories with the bank's personality.
+
+        WHEN TO USE THIS TOOL:
+        Use reflect when you need reasoned analysis, not just fact retrieval. This tool
+        thinks through the question using everything the bank knows and its personality traits.
+
+        EXAMPLES OF GOOD QUERIES:
+        - "What patterns have emerged in how I approach debugging?"
+        - "Based on my past decisions, what architectural style do I prefer?"
+        - "What might be the best approach for this problem given what you know about me?"
+        - "How should I prioritize these tasks based on my goals?"
+
+        HOW IT DIFFERS FROM RECALL:
+        - recall: Returns raw facts matching your search (fast lookup)
+        - reflect: Reasons across memories to form a synthesized answer (deeper analysis)
+
+        Use recall for "what did I say about X?" and reflect for "what should I do about X?"
+
+        Args:
+            query: The question or topic to reflect on
+            context: Optional context about why this reflection is needed
+            budget: Search budget - 'low', 'mid', or 'high' (default: 'low')
+            bank_id: Optional bank to reflect in (defaults to session bank). Use for cross-bank operations.
+        """
+        try:
+            target_bank = bank_id or get_current_bank_id()
+            if target_bank is None:
+                return "Error: No bank_id configured"
+            from hindsight_api.engine.memory_engine import Budget
+
+            # Map string budget to enum
+            budget_map = {"low": Budget.LOW, "mid": Budget.MID, "high": Budget.HIGH}
+            budget_enum = budget_map.get(budget.lower(), Budget.LOW)
+
+            reflect_result = await memory.reflect_async(
+                bank_id=target_bank,
+                query=query,
+                budget=budget_enum,
+                context=context,
+                request_context=RequestContext(),
+            )
+
+            return reflect_result.model_dump_json(indent=2)
+        except Exception as e:
+            logger.error(f"Error reflecting: {e}", exc_info=True)
+            return f'{{"error": "{e}", "text": ""}}'
+
+    @mcp.tool()
+    async def list_banks() -> str:
+        """
+        List all available memory banks.
+
+        Use this tool to discover what memory banks exist in the system.
+        Each bank is an isolated memory store (like a separate "brain").
+
+        Returns:
+            JSON list of banks with their IDs, names, dispositions, and backgrounds.
+        """
+        try:
+            banks = await memory.list_banks(request_context=RequestContext())
+            return json.dumps({"banks": banks}, indent=2)
+        except Exception as e:
+            logger.error(f"Error listing banks: {e}", exc_info=True)
+            return f'{{"error": "{e}", "banks": []}}'
+
+    @mcp.tool()
+    async def create_bank(bank_id: str, name: str | None = None, background: str | None = None) -> str:
+        """
+        Create a new memory bank or get an existing one.
+
+        Memory banks are isolated stores - each one is like a separate "brain" for a user/agent.
+        Banks are auto-created with default settings if they don't exist.
+
+        Args:
+            bank_id: Unique identifier for the bank (e.g., 'user-123', 'agent-alpha')
+            name: Optional human-friendly name for the bank
+            background: Optional background context about the bank's owner/purpose
+        """
+        try:
+            # get_bank_profile auto-creates bank if it doesn't exist
+            profile = await memory.get_bank_profile(bank_id, request_context=RequestContext())
+
+            # Update name/background if provided
+            if name is not None or background is not None:
+                await memory.update_bank(
+                    bank_id,
+                    name=name,
+                    background=background,
+                    request_context=RequestContext(),
+                )
+                # Fetch updated profile
+                profile = await memory.get_bank_profile(bank_id, request_context=RequestContext())
+
+            # Serialize disposition if it's a Pydantic model
+            if "disposition" in profile and hasattr(profile["disposition"], "model_dump"):
+                profile["disposition"] = profile["disposition"].model_dump()
+            return json.dumps(profile, indent=2)
+        except Exception as e:
+            logger.error(f"Error creating bank: {e}", exc_info=True)
+            return f'{{"error": "{e}"}}'

    return mcp

@@ -4,12 +4,9 @@ Centralized configuration for Hindsight API.
 All environment variables and their defaults are defined here.
 """

-import json
 import logging
 import os
-import sys
 from dataclasses import dataclass
-from datetime import datetime, timezone

 from dotenv import find_dotenv, load_dotenv

@@ -39,29 +36,15 @@ ENV_REFLECT_LLM_API_KEY = "HINDSIGHT_API_REFLECT_LLM_API_KEY"
 ENV_REFLECT_LLM_MODEL = "HINDSIGHT_API_REFLECT_LLM_MODEL"
 ENV_REFLECT_LLM_BASE_URL = "HINDSIGHT_API_REFLECT_LLM_BASE_URL"

-ENV_CONSOLIDATION_LLM_PROVIDER = "HINDSIGHT_API_CONSOLIDATION_LLM_PROVIDER"
-ENV_CONSOLIDATION_LLM_API_KEY = "HINDSIGHT_API_CONSOLIDATION_LLM_API_KEY"
-ENV_CONSOLIDATION_LLM_MODEL = "HINDSIGHT_API_CONSOLIDATION_LLM_MODEL"
-ENV_CONSOLIDATION_LLM_BASE_URL = "HINDSIGHT_API_CONSOLIDATION_LLM_BASE_URL"
-
 ENV_EMBEDDINGS_PROVIDER = "HINDSIGHT_API_EMBEDDINGS_PROVIDER"
 ENV_EMBEDDINGS_LOCAL_MODEL = "HINDSIGHT_API_EMBEDDINGS_LOCAL_MODEL"
 ENV_EMBEDDINGS_TEI_URL = "HINDSIGHT_API_EMBEDDINGS_TEI_URL"
 ENV_EMBEDDINGS_OPENAI_API_KEY = "HINDSIGHT_API_EMBEDDINGS_OPENAI_API_KEY"
 ENV_EMBEDDINGS_OPENAI_MODEL = "HINDSIGHT_API_EMBEDDINGS_OPENAI_MODEL"
-ENV_EMBEDDINGS_OPENAI_BASE_URL = "HINDSIGHT_API_EMBEDDINGS_OPENAI_BASE_URL"

 ENV_COHERE_API_KEY = "HINDSIGHT_API_COHERE_API_KEY"
 ENV_EMBEDDINGS_COHERE_MODEL = "HINDSIGHT_API_EMBEDDINGS_COHERE_MODEL"
-ENV_EMBEDDINGS_COHERE_BASE_URL = "HINDSIGHT_API_EMBEDDINGS_COHERE_BASE_URL"
 ENV_RERANKER_COHERE_MODEL = "HINDSIGHT_API_RERANKER_COHERE_MODEL"
-ENV_RERANKER_COHERE_BASE_URL = "HINDSIGHT_API_RERANKER_COHERE_BASE_URL"
-
-# LiteLLM gateway configuration (for embeddings and reranker via LiteLLM proxy)
-ENV_LITELLM_API_BASE = "HINDSIGHT_API_LITELLM_API_BASE"
-ENV_LITELLM_API_KEY = "HINDSIGHT_API_LITELLM_API_KEY"
-ENV_EMBEDDINGS_LITELLM_MODEL = "HINDSIGHT_API_EMBEDDINGS_LITELLM_MODEL"
-ENV_RERANKER_LITELLM_MODEL = "HINDSIGHT_API_RERANKER_LITELLM_MODEL"

 ENV_RERANKER_PROVIDER = "HINDSIGHT_API_RERANKER_PROVIDER"
 ENV_RERANKER_LOCAL_MODEL = "HINDSIGHT_API_RERANKER_LOCAL_MODEL"
@@ -76,7 +59,6 @@ ENV_RERANKER_FLASHRANK_CACHE_DIR = "HINDSIGHT_API_RERANKER_FLASHRANK_CACHE_DIR"
 ENV_HOST = "HINDSIGHT_API_HOST"
 ENV_PORT = "HINDSIGHT_API_PORT"
 ENV_LOG_LEVEL = "HINDSIGHT_API_LOG_LEVEL"
-ENV_LOG_FORMAT = "HINDSIGHT_API_LOG_FORMAT"
 ENV_WORKERS = "HINDSIGHT_API_WORKERS"
 ENV_MCP_ENABLED = "HINDSIGHT_API_MCP_ENABLED"
 ENV_GRAPH_RETRIEVER = "HINDSIGHT_API_GRAPH_RETRIEVER"
@@ -85,20 +67,18 @@ ENV_RECALL_MAX_CONCURRENT = "HINDSIGHT_API_RECALL_MAX_CONCURRENT"
 ENV_RECALL_CONNECTION_BUDGET = "HINDSIGHT_API_RECALL_CONNECTION_BUDGET"
 ENV_MCP_LOCAL_BANK_ID = "HINDSIGHT_API_MCP_LOCAL_BANK_ID"
 ENV_MCP_INSTRUCTIONS = "HINDSIGHT_API_MCP_INSTRUCTIONS"
-ENV_MENTAL_MODEL_REFRESH_CONCURRENCY = "HINDSIGHT_API_MENTAL_MODEL_REFRESH_CONCURRENCY"
+
+# Observation thresholds
+ENV_OBSERVATION_MIN_FACTS = "HINDSIGHT_API_OBSERVATION_MIN_FACTS"
+ENV_OBSERVATION_TOP_ENTITIES = "HINDSIGHT_API_OBSERVATION_TOP_ENTITIES"

 # Retain settings
 ENV_RETAIN_MAX_COMPLETION_TOKENS = "HINDSIGHT_API_RETAIN_MAX_COMPLETION_TOKENS"
 ENV_RETAIN_CHUNK_SIZE = "HINDSIGHT_API_RETAIN_CHUNK_SIZE"
 ENV_RETAIN_EXTRACT_CAUSAL_LINKS = "HINDSIGHT_API_RETAIN_EXTRACT_CAUSAL_LINKS"
 ENV_RETAIN_EXTRACTION_MODE = "HINDSIGHT_API_RETAIN_EXTRACTION_MODE"
-ENV_RETAIN_CUSTOM_INSTRUCTIONS = "HINDSIGHT_API_RETAIN_CUSTOM_INSTRUCTIONS"
 ENV_RETAIN_OBSERVATIONS_ASYNC = "HINDSIGHT_API_RETAIN_OBSERVATIONS_ASYNC"

-# Observations settings (consolidated knowledge from facts)
-ENV_ENABLE_OBSERVATIONS = "HINDSIGHT_API_ENABLE_OBSERVATIONS"
-ENV_CONSOLIDATION_BATCH_SIZE = "HINDSIGHT_API_CONSOLIDATION_BATCH_SIZE"
-
 # Optimization flags
 ENV_SKIP_LLM_VERIFICATION = "HINDSIGHT_API_SKIP_LLM_VERIFICATION"
 ENV_LAZY_RERANKER = "HINDSIGHT_API_LAZY_RERANKER"
@@ -112,16 +92,10 @@ ENV_DB_POOL_MAX_SIZE = "HINDSIGHT_API_DB_POOL_MAX_SIZE"
 ENV_DB_COMMAND_TIMEOUT = "HINDSIGHT_API_DB_COMMAND_TIMEOUT"
 ENV_DB_ACQUIRE_TIMEOUT = "HINDSIGHT_API_DB_ACQUIRE_TIMEOUT"

-# Worker configuration (distributed task processing)
-ENV_WORKER_ENABLED = "HINDSIGHT_API_WORKER_ENABLED"
-ENV_WORKER_ID = "HINDSIGHT_API_WORKER_ID"
-ENV_WORKER_POLL_INTERVAL_MS = "HINDSIGHT_API_WORKER_POLL_INTERVAL_MS"
-ENV_WORKER_MAX_RETRIES = "HINDSIGHT_API_WORKER_MAX_RETRIES"
-ENV_WORKER_BATCH_SIZE = "HINDSIGHT_API_WORKER_BATCH_SIZE"
-ENV_WORKER_HTTP_PORT = "HINDSIGHT_API_WORKER_HTTP_PORT"
-
-# Reflect agent settings
-ENV_REFLECT_MAX_ITERATIONS = "HINDSIGHT_API_REFLECT_MAX_ITERATIONS"
+# Background task processing
+ENV_TASK_BACKEND = "HINDSIGHT_API_TASK_BACKEND"
+ENV_TASK_BACKEND_MEMORY_BATCH_SIZE = "HINDSIGHT_API_TASK_BACKEND_MEMORY_BATCH_SIZE"
+ENV_TASK_BACKEND_MEMORY_BATCH_INTERVAL = "HINDSIGHT_API_TASK_BACKEND_MEMORY_BATCH_INTERVAL"

 # Default values
 DEFAULT_DATABASE_URL = "pg0"
@@ -147,15 +121,9 @@ DEFAULT_RERANKER_FLASHRANK_CACHE_DIR = None  # Use default cache directory
 DEFAULT_EMBEDDINGS_COHERE_MODEL = "embed-english-v3.0"
 DEFAULT_RERANKER_COHERE_MODEL = "rerank-english-v3.0"

-# LiteLLM defaults
-DEFAULT_LITELLM_API_BASE = "http://localhost:4000"
-DEFAULT_EMBEDDINGS_LITELLM_MODEL = "text-embedding-3-small"
-DEFAULT_RERANKER_LITELLM_MODEL = "cohere/rerank-english-v3.0"
-
 DEFAULT_HOST = "0.0.0.0"
 DEFAULT_PORT = 8888
 DEFAULT_LOG_LEVEL = "info"
-DEFAULT_LOG_FORMAT = "text"  # Options: "text", "json"
 DEFAULT_WORKERS = 1
 DEFAULT_MCP_ENABLED = True
 DEFAULT_GRAPH_RETRIEVER = "link_expansion"  # Options: "link_expansion", "mpfp", "bfs"
@@ -163,21 +131,19 @@ DEFAULT_MPFP_TOP_K_NEIGHBORS = 20  # Fan-out limit per node in MPFP graph traver
 DEFAULT_RECALL_MAX_CONCURRENT = 32  # Max concurrent recall operations per worker
 DEFAULT_RECALL_CONNECTION_BUDGET = 4  # Max concurrent DB connections per recall operation
 DEFAULT_MCP_LOCAL_BANK_ID = "mcp"
-DEFAULT_MENTAL_MODEL_REFRESH_CONCURRENCY = 8  # Max concurrent mental model refreshes
+
+# Observation thresholds
+DEFAULT_OBSERVATION_MIN_FACTS = 5  # Min facts required to generate entity observations
+DEFAULT_OBSERVATION_TOP_ENTITIES = 5  # Max entities to process per retain batch

 # Retain settings
 DEFAULT_RETAIN_MAX_COMPLETION_TOKENS = 64000  # Max tokens for fact extraction LLM call
 DEFAULT_RETAIN_CHUNK_SIZE = 3000  # Max chars per chunk for fact extraction
 DEFAULT_RETAIN_EXTRACT_CAUSAL_LINKS = True  # Extract causal links between facts
-DEFAULT_RETAIN_EXTRACTION_MODE = "concise"  # Extraction mode: "concise", "verbose", or "custom"
-RETAIN_EXTRACTION_MODES = ("concise", "verbose", "custom")  # Allowed extraction modes
-DEFAULT_RETAIN_CUSTOM_INSTRUCTIONS = None  # Custom extraction guidelines (only used when mode="custom")
+DEFAULT_RETAIN_EXTRACTION_MODE = "concise"  # Extraction mode: "concise" or "verbose"
+RETAIN_EXTRACTION_MODES = ("concise", "verbose")  # Allowed extraction modes
 DEFAULT_RETAIN_OBSERVATIONS_ASYNC = False  # Run observation generation async (after retain completes)

-# Observations defaults (consolidated knowledge from facts)
-DEFAULT_ENABLE_OBSERVATIONS = True  # Observations enabled by default
-DEFAULT_CONSOLIDATION_BATCH_SIZE = 50  # Memories to load per batch (internal memory optimization)
-
 # Database migrations
 DEFAULT_RUN_MIGRATIONS_ON_STARTUP = True

@@ -187,16 +153,10 @@ DEFAULT_DB_POOL_MAX_SIZE = 100
 DEFAULT_DB_COMMAND_TIMEOUT = 60  # seconds
 DEFAULT_DB_ACQUIRE_TIMEOUT = 30  # seconds

-# Worker configuration (distributed task processing)
-DEFAULT_WORKER_ENABLED = True  # API runs worker by default (standalone mode)
-DEFAULT_WORKER_ID = None  # Will use hostname if not specified
-DEFAULT_WORKER_POLL_INTERVAL_MS = 500  # Poll database every 500ms
-DEFAULT_WORKER_MAX_RETRIES = 3  # Max retries before marking task failed
-DEFAULT_WORKER_BATCH_SIZE = 10  # Tasks to claim per poll cycle
-DEFAULT_WORKER_HTTP_PORT = 8889  # HTTP port for worker metrics/health
-
-# Reflect agent settings
-DEFAULT_REFLECT_MAX_ITERATIONS = 10  # Max tool call iterations before forcing response
+# Background task processing
+DEFAULT_TASK_BACKEND = "memory"  # Options: "memory", "noop"
+DEFAULT_TASK_BACKEND_MEMORY_BATCH_SIZE = 10
+DEFAULT_TASK_BACKEND_MEMORY_BATCH_INTERVAL = 1.0  # seconds

 # Default MCP tool descriptions (can be customized via env vars)
 DEFAULT_MCP_RETAIN_DESCRIPTION = """Store important information to long-term memory.
@@ -222,36 +182,6 @@ Use this tool PROACTIVELY to:
 EMBEDDING_DIMENSION = DEFAULT_EMBEDDING_DIMENSION


-class JsonFormatter(logging.Formatter):
-    """JSON formatter for structured logging.
-
-    Outputs logs in JSON format with a 'severity' field that cloud logging
-    systems (GCP, AWS CloudWatch, etc.) can parse to correctly categorize log levels.
-    """
-
-    SEVERITY_MAP = {
-        logging.DEBUG: "DEBUG",
-        logging.INFO: "INFO",
-        logging.WARNING: "WARNING",
-        logging.ERROR: "ERROR",
-        logging.CRITICAL: "CRITICAL",
-    }
-
-    def format(self, record: logging.LogRecord) -> str:
-        log_entry = {
-            "severity": self.SEVERITY_MAP.get(record.levelno, "DEFAULT"),
-            "message": record.getMessage(),
-            "timestamp": datetime.now(timezone.utc).isoformat(),
-            "logger": record.name,
-        }
-
-        # Add exception info if present
-        if record.exc_info:
-            log_entry["exception"] = self.formatException(record.exc_info)
-
-        return json.dumps(log_entry)
-
-
 def _validate_extraction_mode(mode: str) -> str:
    """Validate and normalize extraction mode."""
    mode_lower = mode.lower()
@@ -290,17 +220,10 @@ class HindsightConfig:
    reflect_llm_model: str | None
    reflect_llm_base_url: str | None

-    consolidation_llm_provider: str | None
-    consolidation_llm_api_key: str | None
-    consolidation_llm_model: str | None
-    consolidation_llm_base_url: str | None
-
    # Embeddings
    embeddings_provider: str
    embeddings_local_model: str
    embeddings_tei_url: str | None
-    embeddings_openai_base_url: str | None
-    embeddings_cohere_base_url: str | None

    # Reranker
    reranker_provider: str
@@ -309,13 +232,11 @@ class HindsightConfig:
    reranker_tei_batch_size: int
    reranker_tei_max_concurrent: int
    reranker_max_candidates: int
-    reranker_cohere_base_url: str | None

    # Server
    host: str
    port: int
    log_level: str
-    log_format: str
    mcp_enabled: bool

    # Recall
@@ -323,20 +244,18 @@ class HindsightConfig:
    mpfp_top_k_neighbors: int
    recall_max_concurrent: int
    recall_connection_budget: int
-    mental_model_refresh_concurrency: int
+
+    # Observation thresholds
+    observation_min_facts: int
+    observation_top_entities: int

    # Retain settings
    retain_max_completion_tokens: int
    retain_chunk_size: int
    retain_extract_causal_links: bool
    retain_extraction_mode: str
-    retain_custom_instructions: str | None
    retain_observations_async: bool

-    # Observations settings (consolidated knowledge from facts)
-    enable_observations: bool
-    consolidation_batch_size: int
-
    # Optimization flags
    skip_llm_verification: bool
    lazy_reranker: bool
@@ -350,16 +269,10 @@ class HindsightConfig:
    db_command_timeout: int
    db_acquire_timeout: int

-    # Worker configuration (distributed task processing)
-    worker_enabled: bool
-    worker_id: str | None
-    worker_poll_interval_ms: int
-    worker_max_retries: int
-    worker_batch_size: int
-    worker_http_port: int
-
-    # Reflect agent settings
-    reflect_max_iterations: int
+    # Background task processing
+    task_backend: str
+    task_backend_memory_batch_size: int
+    task_backend_memory_batch_interval: float

    @classmethod
    def from_env(cls) -> "HindsightConfig":
@@ -383,16 +296,10 @@ class HindsightConfig:
            reflect_llm_api_key=os.getenv(ENV_REFLECT_LLM_API_KEY) or None,
            reflect_llm_model=os.getenv(ENV_REFLECT_LLM_MODEL) or None,
            reflect_llm_base_url=os.getenv(ENV_REFLECT_LLM_BASE_URL) or None,
-            consolidation_llm_provider=os.getenv(ENV_CONSOLIDATION_LLM_PROVIDER) or None,
-            consolidation_llm_api_key=os.getenv(ENV_CONSOLIDATION_LLM_API_KEY) or None,
-            consolidation_llm_model=os.getenv(ENV_CONSOLIDATION_LLM_MODEL) or None,
-            consolidation_llm_base_url=os.getenv(ENV_CONSOLIDATION_LLM_BASE_URL) or None,
            # Embeddings
            embeddings_provider=os.getenv(ENV_EMBEDDINGS_PROVIDER, DEFAULT_EMBEDDINGS_PROVIDER),
            embeddings_local_model=os.getenv(ENV_EMBEDDINGS_LOCAL_MODEL, DEFAULT_EMBEDDINGS_LOCAL_MODEL),
            embeddings_tei_url=os.getenv(ENV_EMBEDDINGS_TEI_URL),
-            embeddings_openai_base_url=os.getenv(ENV_EMBEDDINGS_OPENAI_BASE_URL) or None,
-            embeddings_cohere_base_url=os.getenv(ENV_EMBEDDINGS_COHERE_BASE_URL) or None,
            # Reranker
            reranker_provider=os.getenv(ENV_RERANKER_PROVIDER, DEFAULT_RERANKER_PROVIDER),
            reranker_local_model=os.getenv(ENV_RERANKER_LOCAL_MODEL, DEFAULT_RERANKER_LOCAL_MODEL),
@@ -402,12 +309,10 @@ class HindsightConfig:
                os.getenv(ENV_RERANKER_TEI_MAX_CONCURRENT, str(DEFAULT_RERANKER_TEI_MAX_CONCURRENT))
            ),
            reranker_max_candidates=int(os.getenv(ENV_RERANKER_MAX_CANDIDATES, str(DEFAULT_RERANKER_MAX_CANDIDATES))),
-            reranker_cohere_base_url=os.getenv(ENV_RERANKER_COHERE_BASE_URL) or None,
            # Server
            host=os.getenv(ENV_HOST, DEFAULT_HOST),
            port=int(os.getenv(ENV_PORT, DEFAULT_PORT)),
            log_level=os.getenv(ENV_LOG_LEVEL, DEFAULT_LOG_LEVEL),
-            log_format=os.getenv(ENV_LOG_FORMAT, DEFAULT_LOG_FORMAT).lower(),
            mcp_enabled=os.getenv(ENV_MCP_ENABLED, str(DEFAULT_MCP_ENABLED)).lower() == "true",
            # Recall
            graph_retriever=os.getenv(ENV_GRAPH_RETRIEVER, DEFAULT_GRAPH_RETRIEVER),
@@ -416,12 +321,14 @@ class HindsightConfig:
            recall_connection_budget=int(
                os.getenv(ENV_RECALL_CONNECTION_BUDGET, str(DEFAULT_RECALL_CONNECTION_BUDGET))
            ),
-            mental_model_refresh_concurrency=int(
-                os.getenv(ENV_MENTAL_MODEL_REFRESH_CONCURRENCY, str(DEFAULT_MENTAL_MODEL_REFRESH_CONCURRENCY))
-            ),
            # Optimization flags
            skip_llm_verification=os.getenv(ENV_SKIP_LLM_VERIFICATION, "false").lower() == "true",
            lazy_reranker=os.getenv(ENV_LAZY_RERANKER, "false").lower() == "true",
+            # Observation thresholds
+            observation_min_facts=int(os.getenv(ENV_OBSERVATION_MIN_FACTS, str(DEFAULT_OBSERVATION_MIN_FACTS))),
+            observation_top_entities=int(
+                os.getenv(ENV_OBSERVATION_TOP_ENTITIES, str(DEFAULT_OBSERVATION_TOP_ENTITIES))
+            ),
            # Retain settings
            retain_max_completion_tokens=int(
                os.getenv(ENV_RETAIN_MAX_COMPLETION_TOKENS, str(DEFAULT_RETAIN_MAX_COMPLETION_TOKENS))
@@ -434,16 +341,10 @@ class HindsightConfig:
            retain_extraction_mode=_validate_extraction_mode(
                os.getenv(ENV_RETAIN_EXTRACTION_MODE, DEFAULT_RETAIN_EXTRACTION_MODE)
            ),
-            retain_custom_instructions=os.getenv(ENV_RETAIN_CUSTOM_INSTRUCTIONS) or DEFAULT_RETAIN_CUSTOM_INSTRUCTIONS,
            retain_observations_async=os.getenv(
                ENV_RETAIN_OBSERVATIONS_ASYNC, str(DEFAULT_RETAIN_OBSERVATIONS_ASYNC)
            ).lower()
            == "true",
-            # Observations settings (consolidated knowledge from facts)
-            enable_observations=os.getenv(ENV_ENABLE_OBSERVATIONS, str(DEFAULT_ENABLE_OBSERVATIONS)).lower() == "true",
-            consolidation_batch_size=int(
-                os.getenv(ENV_CONSOLIDATION_BATCH_SIZE, str(DEFAULT_CONSOLIDATION_BATCH_SIZE))
-            ),
            # Database migrations
            run_migrations_on_startup=os.getenv(ENV_RUN_MIGRATIONS_ON_STARTUP, "true").lower() == "true",
            # Database connection pool
@@ -451,15 +352,14 @@ class HindsightConfig:
            db_pool_max_size=int(os.getenv(ENV_DB_POOL_MAX_SIZE, str(DEFAULT_DB_POOL_MAX_SIZE))),
            db_command_timeout=int(os.getenv(ENV_DB_COMMAND_TIMEOUT, str(DEFAULT_DB_COMMAND_TIMEOUT))),
            db_acquire_timeout=int(os.getenv(ENV_DB_ACQUIRE_TIMEOUT, str(DEFAULT_DB_ACQUIRE_TIMEOUT))),
-            # Worker configuration
-            worker_enabled=os.getenv(ENV_WORKER_ENABLED, str(DEFAULT_WORKER_ENABLED)).lower() == "true",
-            worker_id=os.getenv(ENV_WORKER_ID) or DEFAULT_WORKER_ID,
-            worker_poll_interval_ms=int(os.getenv(ENV_WORKER_POLL_INTERVAL_MS, str(DEFAULT_WORKER_POLL_INTERVAL_MS))),
-            worker_max_retries=int(os.getenv(ENV_WORKER_MAX_RETRIES, str(DEFAULT_WORKER_MAX_RETRIES))),
-            worker_batch_size=int(os.getenv(ENV_WORKER_BATCH_SIZE, str(DEFAULT_WORKER_BATCH_SIZE))),
-            worker_http_port=int(os.getenv(ENV_WORKER_HTTP_PORT, str(DEFAULT_WORKER_HTTP_PORT))),
-            # Reflect agent settings
-            reflect_max_iterations=int(os.getenv(ENV_REFLECT_MAX_ITERATIONS, str(DEFAULT_REFLECT_MAX_ITERATIONS))),
+            # Background task processing
+            task_backend=os.getenv(ENV_TASK_BACKEND, DEFAULT_TASK_BACKEND),
+            task_backend_memory_batch_size=int(
+                os.getenv(ENV_TASK_BACKEND_MEMORY_BATCH_SIZE, str(DEFAULT_TASK_BACKEND_MEMORY_BATCH_SIZE))
+            ),
+            task_backend_memory_batch_interval=float(
+                os.getenv(ENV_TASK_BACKEND_MEMORY_BATCH_INTERVAL, str(DEFAULT_TASK_BACKEND_MEMORY_BATCH_INTERVAL))
+            ),
        )

    def get_llm_base_url(self) -> str:
@@ -490,28 +390,12 @@ class HindsightConfig:
        return log_level_map.get(self.log_level.lower(), logging.INFO)

    def configure_logging(self) -> None:
-        """Configure Python logging based on the log level and format.
-
-        When log_format is "json", outputs structured JSON logs with a severity
-        field that GCP Cloud Logging can parse for proper log level categorization.
-        """
-        root_logger = logging.getLogger()
-        root_logger.setLevel(self.get_python_log_level())
-
-        # Remove existing handlers
-        for handler in root_logger.handlers[:]:
-            root_logger.removeHandler(handler)
-
-        # Create handler writing to stdout (GCP treats stderr as ERROR)
-        handler = logging.StreamHandler(sys.stdout)
-        handler.setLevel(self.get_python_log_level())
-
-        if self.log_format == "json":
-            handler.setFormatter(JsonFormatter())
-        else:
-            handler.setFormatter(logging.Formatter("%(asctime)s - %(levelname)s - %(name)s - %(message)s"))
-
-        root_logger.addHandler(handler)
+        """Configure Python logging based on the log level."""
+        logging.basicConfig(
+            level=self.get_python_log_level(),
+            format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
+            force=True,  # Override any existing configuration
+        )

    def log_config(self) -> None:
        """Log the current configuration (without sensitive values)."""
@@ -525,10 +409,6 @@ class HindsightConfig:
            reflect_provider = self.reflect_llm_provider or self.llm_provider
            reflect_model = self.reflect_llm_model or self.llm_model
            logger.info(f"LLM (reflect): provider={reflect_provider}, model={reflect_model}")
-        if self.consolidation_llm_provider or self.consolidation_llm_model:
-            consolidation_provider = self.consolidation_llm_provider or self.llm_provider
-            consolidation_model = self.consolidation_llm_model or self.llm_model
-            logger.info(f"LLM (consolidation): provider={consolidation_provider}, model={consolidation_model}")
        logger.info(f"Embeddings: provider={self.embeddings_provider}")
        logger.info(f"Reranker: provider={self.reranker_provider}")
        logger.info(f"Graph retriever: {self.graph_retriever}")
@@ -1,5 +0,0 @@
-"""Consolidation engine for automatic learning creation from memories."""
-
-from .consolidator import run_consolidation_job
-
-__all__ = ["run_consolidation_job"]
@@ -1,859 +0,0 @@
-"""Consolidation engine for automatic observation creation from memories.
-
-The consolidation engine runs as a background job after retain operations complete.
-It processes new memories and either:
- Creates new observations from novel facts
- Updates existing observations when new evidence supports/contradicts/refines them
-
-Observations are stored in memory_units with fact_type='observation' and include:
- proof_count: Number of supporting memories
- source_memory_ids: Array of memory UUIDs that contribute to this observation
- history: JSONB tracking changes over time
-"""
-
-import json
-import logging
-import time
-import uuid
-from datetime import datetime, timezone
-from typing import TYPE_CHECKING, Any
-
-from ..memory_engine import fq_table
-from ..retain import embedding_utils
-from .prompts import (
-    CONSOLIDATION_SYSTEM_PROMPT,
-    CONSOLIDATION_USER_PROMPT,
-)
-
-if TYPE_CHECKING:
-    from asyncpg import Connection
-
-    from ...api.http import RequestContext
-    from ..memory_engine import MemoryEngine
-
-logger = logging.getLogger(__name__)
-
-
-class ConsolidationPerfLog:
-    """Performance logging for consolidation operations."""
-
-    def __init__(self, bank_id: str):
-        self.bank_id = bank_id
-        self.start_time = time.time()
-        self.lines: list[str] = []
-        self.timings: dict[str, float] = {}
-
-    def log(self, message: str) -> None:
-        """Add a log line."""
-        self.lines.append(message)
-
-    def record_timing(self, key: str, duration: float) -> None:
-        """Record a timing measurement."""
-        if key in self.timings:
-            self.timings[key] += duration
-        else:
-            self.timings[key] = duration
-
-    def flush(self) -> None:
-        """Flush all log lines to the logger."""
-        total_time = time.time() - self.start_time
-        header = f"\n{'=' * 60}\nCONSOLIDATION for bank {self.bank_id}"
-        footer = f"{'=' * 60}\nCONSOLIDATION COMPLETE: {total_time:.3f}s total\n{'=' * 60}"
-
-        log_output = header + "\n" + "\n".join(self.lines) + "\n" + footer
-        logger.info(log_output)
-
-
-async def run_consolidation_job(
-    memory_engine: "MemoryEngine",
-    bank_id: str,
-    request_context: "RequestContext",
-) -> dict[str, Any]:
-    """
-    Run consolidation job for a bank.
-
-    This is called after retain operations to consolidate new memories into mental models.
-
-    Args:
-        memory_engine: MemoryEngine instance
-        bank_id: Bank identifier
-        request_context: Request context for authentication
-
-    Returns:
-        Dict with consolidation results
-    """
-    from ...config import get_config
-
-    config = get_config()
-    perf = ConsolidationPerfLog(bank_id)
-    max_memories_per_batch = config.consolidation_batch_size
-
-    # Check if consolidation is enabled
-    if not config.enable_observations:
-        logger.debug(f"Consolidation disabled for bank {bank_id}")
-        return {"status": "disabled", "bank_id": bank_id}
-
-    pool = memory_engine._pool
-
-    # Get bank profile
-    async with pool.acquire() as conn:
-        t0 = time.time()
-        bank_row = await conn.fetchrow(
-            f"""
-            SELECT bank_id, name, mission
-            FROM {fq_table("banks")}
-            WHERE bank_id = $1
-            """,
-            bank_id,
-        )
-
-        if not bank_row:
-            logger.warning(f"Bank {bank_id} not found for consolidation")
-            return {"status": "bank_not_found", "bank_id": bank_id}
-
-        mission = bank_row["mission"] or "General memory consolidation"
-        perf.record_timing("fetch_bank", time.time() - t0)
-
-        # Count total unconsolidated memories for progress logging
-        total_count = await conn.fetchval(
-            f"""
-            SELECT COUNT(*)
-            FROM {fq_table("memory_units")}
-            WHERE bank_id = $1
-              AND consolidated_at IS NULL
-              AND fact_type IN ('experience', 'world')
-            """,
-            bank_id,
-        )
-
-    if total_count == 0:
-        logger.debug(f"No new memories to consolidate for bank {bank_id}")
-        return {"status": "no_new_memories", "bank_id": bank_id, "memories_processed": 0}
-
-    logger.info(f"[CONSOLIDATION] bank={bank_id} total_unconsolidated={total_count}")
-    perf.log(f"[1] Found {total_count} pending memories to consolidate")
-
-    # Process each memory with individual commits for crash recovery
-    stats = {
-        "memories_processed": 0,
-        "observations_created": 0,
-        "observations_updated": 0,
-        "observations_merged": 0,
-        "actions_executed": 0,
-        "skipped": 0,
-    }
-
-    batch_num = 0
-    while True:
-        batch_num += 1
-        batch_start = time.time()
-
-        # Fetch next batch of unconsolidated memories
-        async with pool.acquire() as conn:
-            t0 = time.time()
-            memories = await conn.fetch(
-                f"""
-                SELECT id, text, fact_type, occurred_start, occurred_end, event_date, tags, mentioned_at
-                FROM {fq_table("memory_units")}
-                WHERE bank_id = $1
-                  AND consolidated_at IS NULL
-                  AND fact_type IN ('experience', 'world')
-                ORDER BY created_at ASC
-                LIMIT $2
-                """,
-                bank_id,
-                max_memories_per_batch,
-            )
-            perf.record_timing("fetch_memories", time.time() - t0)
-
-        if not memories:
-            break  # No more unconsolidated memories
-
-        for memory in memories:
-            mem_start = time.time()
-
-            # Process the memory (uses its own connection internally)
-            async with pool.acquire() as conn:
-                result = await _process_memory(
-                    conn=conn,
-                    memory_engine=memory_engine,
-                    bank_id=bank_id,
-                    memory=dict(memory),
-                    mission=mission,
-                    request_context=request_context,
-                    perf=perf,
-                )
-
-                # Mark memory as consolidated (committed immediately)
-                await conn.execute(
-                    f"""
-                    UPDATE {fq_table("memory_units")}
-                    SET consolidated_at = NOW()
-                    WHERE id = $1
-                    """,
-                    memory["id"],
-                )
-
-            mem_time = time.time() - mem_start
-            perf.record_timing("process_memory_total", mem_time)
-
-            stats["memories_processed"] += 1
-
-            action = result.get("action")
-            if action == "created":
-                stats["observations_created"] += 1
-                stats["actions_executed"] += 1
-            elif action == "updated":
-                stats["observations_updated"] += 1
-                stats["actions_executed"] += 1
-            elif action == "merged":
-                stats["observations_merged"] += 1
-                stats["actions_executed"] += 1
-            elif action == "multiple":
-                stats["observations_created"] += result.get("created", 0)
-                stats["observations_updated"] += result.get("updated", 0)
-                stats["observations_merged"] += result.get("merged", 0)
-                stats["actions_executed"] += result.get("total_actions", 0)
-            elif action == "skipped":
-                stats["skipped"] += 1
-
-            # Log progress periodically
-            if stats["memories_processed"] % 10 == 0:
-                logger.info(
-                    f"[CONSOLIDATION] bank={bank_id} progress: "
-                    f"{stats['memories_processed']}/{total_count} memories processed"
-                )
-
-        batch_time = time.time() - batch_start
-        perf.log(
-            f"[2] Batch {batch_num}: {len(memories)} memories in {batch_time:.3f}s "
-            f"(avg {batch_time / len(memories):.3f}s/memory)"
-        )
-
-    # Build summary
-    perf.log(
-        f"[3] Results: {stats['memories_processed']} memories -> "
-        f"{stats['actions_executed']} actions "
-        f"({stats['observations_created']} created, "
-        f"{stats['observations_updated']} updated, "
-        f"{stats['observations_merged']} merged, "
-        f"{stats['skipped']} skipped)"
-    )
-
-    # Add timing breakdown
-    timing_parts = []
-    if "recall" in perf.timings:
-        timing_parts.append(f"recall={perf.timings['recall']:.3f}s")
-    if "llm" in perf.timings:
-        timing_parts.append(f"llm={perf.timings['llm']:.3f}s")
-    if "embedding" in perf.timings:
-        timing_parts.append(f"embedding={perf.timings['embedding']:.3f}s")
-    if "db_write" in perf.timings:
-        timing_parts.append(f"db_write={perf.timings['db_write']:.3f}s")
-
-    if timing_parts:
-        perf.log(f"[4] Timing breakdown: {', '.join(timing_parts)}")
-
-    # Trigger mental model refreshes for models with refresh_after_consolidation=true
-    mental_models_refreshed = await _trigger_mental_model_refreshes(
-        memory_engine=memory_engine,
-        bank_id=bank_id,
-        request_context=request_context,
-        perf=perf,
-    )
-    stats["mental_models_refreshed"] = mental_models_refreshed
-
-    perf.flush()
-
-    return {"status": "completed", "bank_id": bank_id, **stats}
-
-
-async def _trigger_mental_model_refreshes(
-    memory_engine: "MemoryEngine",
-    bank_id: str,
-    request_context: "RequestContext",
-    perf: ConsolidationPerfLog | None = None,
-) -> int:
-    """
-    Trigger refreshes for mental models with refresh_after_consolidation=true.
-
-    Args:
-        memory_engine: MemoryEngine instance
-        bank_id: Bank identifier
-        request_context: Request context for authentication
-        perf: Performance logging
-
-    Returns:
-        Number of mental models scheduled for refresh
-    """
-    pool = memory_engine._pool
-
-    # Find mental models with refresh_after_consolidation=true
-    async with pool.acquire() as conn:
-        rows = await conn.fetch(
-            f"""
-            SELECT id, name
-            FROM {fq_table("mental_models")}
-            WHERE bank_id = $1
-              AND (trigger->>'refresh_after_consolidation')::boolean = true
-            """,
-            bank_id,
-        )
-
-    if not rows:
-        return 0
-
-    if perf:
-        perf.log(f"[5] Triggering refresh for {len(rows)} mental models with refresh_after_consolidation=true")
-
-    # Submit refresh tasks for each mental model
-    refreshed_count = 0
-    for row in rows:
-        mental_model_id = row["id"]
-        try:
-            await memory_engine.submit_async_refresh_mental_model(
-                bank_id=bank_id,
-                mental_model_id=mental_model_id,
-                request_context=request_context,
-            )
-            refreshed_count += 1
-            logger.info(
-                f"[CONSOLIDATION] Triggered refresh for mental model {mental_model_id} "
-                f"(name: {row['name']}) in bank {bank_id}"
-            )
-        except Exception as e:
-            logger.warning(f"[CONSOLIDATION] Failed to trigger refresh for mental model {mental_model_id}: {e}")
-
-    return refreshed_count
-
-
-async def _process_memory(
-    conn: "Connection",
-    memory_engine: "MemoryEngine",
-    bank_id: str,
-    memory: dict[str, Any],
-    mission: str,
-    request_context: "RequestContext",
-    perf: ConsolidationPerfLog | None = None,
-) -> dict[str, Any]:
-    """
-    Process a single memory for consolidation using a SINGLE LLM call.
-
-    This function:
-    1. Finds related observations (can be empty)
-    2. Uses ONE LLM call to extract durable knowledge AND decide on actions
-    3. Executes array of actions (can be multiple creates/updates)
-
-    The LLM handles all cases:
-    - No related observations: returns create action(s) with extracted durable knowledge
-    - Related observations exist: returns update/create actions based on tag routing
-    - Purely ephemeral fact: returns empty array (skip)
-
-    Returns:
-        Dict with action summary: created/updated/merged counts
-    """
-    fact_text = memory["text"]
-    memory_id = memory["id"]
-    fact_tags = memory.get("tags") or []
-
-    # Find related observations using the full recall system (NO tag filtering)
-    t0 = time.time()
-    related_observations = await _find_related_observations(
-        conn=conn,
-        memory_engine=memory_engine,
-        bank_id=bank_id,
-        query=fact_text,
-        request_context=request_context,
-    )
-    if perf:
-        perf.record_timing("recall", time.time() - t0)
-
-    # Single LLM call handles ALL cases (with or without existing observations)
-    # Note: Tags are NOT passed to LLM - they are handled algorithmically
-    t0 = time.time()
-    actions = await _consolidate_with_llm(
-        memory_engine=memory_engine,
-        fact_text=fact_text,
-        observations=related_observations,  # Can be empty list
-        mission=mission,
-    )
-    if perf:
-        perf.record_timing("llm", time.time() - t0)
-
-    if not actions:
-        # LLM returned empty array - fact is purely ephemeral, skip
-        return {"action": "skipped", "reason": "no_durable_knowledge"}
-
-    # Execute all actions and collect results
-    results = []
-    for action in actions:
-        action_type = action.get("action")
-        if action_type == "update":
-            result = await _execute_update_action(
-                conn=conn,
-                memory_engine=memory_engine,
-                bank_id=bank_id,
-                memory_id=memory_id,
-                action=action,
-                observations=related_observations,
-                source_fact_tags=fact_tags,  # Pass source fact's tags for security
-                source_occurred_start=memory.get("occurred_start"),
-                source_occurred_end=memory.get("occurred_end"),
-                source_mentioned_at=memory.get("mentioned_at"),
-                perf=perf,
-            )
-            results.append(result)
-        elif action_type == "create":
-            result = await _execute_create_action(
-                conn=conn,
-                memory_engine=memory_engine,
-                bank_id=bank_id,
-                memory_id=memory_id,
-                action=action,
-                source_fact_tags=fact_tags,  # Pass source fact's tags for security
-                event_date=memory.get("event_date"),
-                occurred_start=memory.get("occurred_start"),
-                occurred_end=memory.get("occurred_end"),
-                mentioned_at=memory.get("mentioned_at"),
-                perf=perf,
-            )
-            results.append(result)
-
-    if not results:
-        # No valid actions executed
-        return {"action": "skipped", "reason": "no_valid_actions"}
-
-    # Summarize results
-    created = sum(1 for r in results if r.get("action") == "created")
-    updated = sum(1 for r in results if r.get("action") == "updated")
-    merged = sum(1 for r in results if r.get("action") == "merged")
-
-    if len(results) == 1:
-        return results[0]
-
-    return {
-        "action": "multiple",
-        "created": created,
-        "updated": updated,
-        "merged": merged,
-        "total_actions": len(results),
-    }
-
-
-async def _execute_update_action(
-    conn: "Connection",
-    memory_engine: "MemoryEngine",
-    bank_id: str,
-    memory_id: uuid.UUID,
-    action: dict[str, Any],
-    observations: list[dict[str, Any]],
-    source_fact_tags: list[str] | None = None,
-    source_occurred_start: datetime | None = None,
-    source_occurred_end: datetime | None = None,
-    source_mentioned_at: datetime | None = None,
-    perf: ConsolidationPerfLog | None = None,
-) -> dict[str, Any]:
-    """
-    Execute an update action on an existing observation.
-
-    Updates the observation text, adds to history, increments proof_count,
-    and updates temporal fields:
-    - occurred_start: uses LEAST to keep the earliest start time
-    - occurred_end: uses GREATEST to keep the most recent end time
-    - mentioned_at: uses GREATEST to keep the most recent mention time
-
-    SECURITY: Merges source fact's tags into the observation's existing tags.
-    This ensures all contributors can see the observation they contributed to.
-    For example, if Lisa's observation (tags=['user_lisa']) is updated with
-    Mike's fact (tags=['user_mike']), the observation will have both tags.
-    """
-    learning_id = action.get("learning_id")
-    new_text = action.get("text")
-    reason = action.get("reason", "Updated with new fact")
-
-    if not learning_id or not new_text:
-        return {"action": "skipped", "reason": "missing_learning_id_or_text"}
-
-    # Find the observation
-    model = next((m for m in observations if str(m["id"]) == learning_id), None)
-    if not model:
-        return {"action": "skipped", "reason": "learning_not_found"}
-
-    # Build history entry
-    history = list(model.get("history", []))
-    history.append(
-        {
-            "previous_text": model["text"],
-            "changed_at": datetime.now(timezone.utc).isoformat(),
-            "reason": reason,
-            "source_memory_id": str(memory_id),
-        }
-    )
-
-    # Update source_memory_ids
-    source_ids = list(model.get("source_memory_ids", []))
-    source_ids.append(memory_id)
-
-    # SECURITY: Merge source fact's tags into existing observation tags
-    # This ensures all contributors can see the observation they contributed to
-    existing_tags = set(model.get("tags", []) or [])
-    source_tags = set(source_fact_tags or [])
-    merged_tags = list(existing_tags | source_tags)  # Union of both tag sets
-    if source_tags and source_tags != existing_tags:
-        logger.debug(
-            f"Security: Merging tags for observation {learning_id}: "
-            f"existing={list(existing_tags)}, source={list(source_tags)}, merged={merged_tags}"
-        )
-
-    # Generate new embedding for updated text
-    t0 = time.time()
-    embeddings = await embedding_utils.generate_embeddings_batch(memory_engine.embeddings, [new_text])
-    embedding_str = str(embeddings[0]) if embeddings else None
-    if perf:
-        perf.record_timing("embedding", time.time() - t0)
-
-    # Update the observation
-    # - occurred_start: LEAST keeps the earliest start time across all source facts
-    # - occurred_end: GREATEST keeps the most recent end time across all source facts
-    # - mentioned_at: GREATEST keeps the most recent mention time
-    # - tags: merged from existing + source fact (for visibility)
-    t0 = time.time()
-    await conn.execute(
-        f"""
-        UPDATE {fq_table("memory_units")}
-        SET text = $1,
-            embedding = $2::vector,
-            history = $3,
-            source_memory_ids = $4,
-            proof_count = $5,
-            tags = $10,
-            updated_at = now(),
-            occurred_start = LEAST(occurred_start, COALESCE($7, occurred_start)),
-            occurred_end = GREATEST(occurred_end, COALESCE($8, occurred_end)),
-            mentioned_at = GREATEST(mentioned_at, COALESCE($9, mentioned_at))
-        WHERE id = $6
-        """,
-        new_text,
-        embedding_str,
-        json.dumps(history),
-        source_ids,
-        len(source_ids),
-        uuid.UUID(learning_id),
-        source_occurred_start,
-        source_occurred_end,
-        source_mentioned_at,
-        merged_tags,
-    )
-
-    # Create links from memory to observation
-    await _create_memory_links(conn, memory_id, uuid.UUID(learning_id))
-    if perf:
-        perf.record_timing("db_write", time.time() - t0)
-
-    logger.debug(f"Updated observation {learning_id} with memory {memory_id}")
-
-    return {"action": "updated", "observation_id": learning_id}
-
-
-async def _execute_create_action(
-    conn: "Connection",
-    memory_engine: "MemoryEngine",
-    bank_id: str,
-    memory_id: uuid.UUID,
-    action: dict[str, Any],
-    source_fact_tags: list[str] | None = None,
-    event_date: datetime | None = None,
-    occurred_start: datetime | None = None,
-    occurred_end: datetime | None = None,
-    mentioned_at: datetime | None = None,
-    perf: ConsolidationPerfLog | None = None,
-) -> dict[str, Any]:
-    """
-    Execute a create action for a new observation.
-
-    Creates a new observation with the specified text.
-    The text comes directly from the classify LLM - no second LLM call needed.
-
-    Tags are determined algorithmically (not by LLM):
-    - Observations always inherit their source fact's tags
-    - This ensures visibility scope is maintained (security)
-    """
-    text = action.get("text")
-
-    # Tags are determined algorithmically - always use source fact's tags
-    # This ensures private memories create private observations
-    tags = source_fact_tags or []
-
-    if not text:
-        return {"action": "skipped", "reason": "missing_text"}
-
-    # Use text directly from classify - skip the redundant LLM call
-    result = await _create_observation_directly(
-        conn=conn,
-        memory_engine=memory_engine,
-        bank_id=bank_id,
-        source_memory_id=memory_id,
-        observation_text=text,  # Text already processed by classify LLM
-        tags=tags,
-        event_date=event_date,
-        occurred_start=occurred_start,
-        occurred_end=occurred_end,
-        mentioned_at=mentioned_at,
-        perf=perf,
-    )
-
-    logger.debug(f"Created observation {result.get('observation_id')} from memory {memory_id} (tags: {tags})")
-
-    return result
-
-
-async def _create_memory_links(
-    conn: "Connection",
-    memory_id: uuid.UUID,
-    observation_id: uuid.UUID,
-) -> None:
-    """
-    Placeholder for observation link creation.
-
-    Observations do NOT get any memory_links copied from their source facts.
-    Instead, retrieval uses source_memory_ids to traverse:
-    - Entity connections: observation → source_memory_ids → unit_entities
-    - Semantic similarity: observations have their own embeddings
-    - Temporal proximity: observations have their own temporal fields
-
-    This avoids data duplication and ensures observations are always
-    connected via their source facts' relationships.
-
-    The memory_id and observation_id parameters are kept for interface
-    compatibility but no links are created.
-    """
-    # No links are created - observations rely on source_memory_ids for traversal
-    pass
-
-
-async def _find_related_observations(
-    conn: "Connection",
-    memory_engine: "MemoryEngine",
-    bank_id: str,
-    query: str,
-    request_context: "RequestContext",
-) -> list[dict[str, Any]]:
-    """
-    Find observations related to the given query using the full recall system.
-
-    IMPORTANT: We do NOT filter by tags here. Consolidation needs to see ALL
-    potentially related observations regardless of scope, so the LLM can
-    decide on tag routing (same scope update vs cross-scope create).
-
-    This leverages:
-    - Semantic search (embedding similarity)
-    - BM25 text search (keyword matching)
-    - Entity-based retrieval (shared entities)
-    - Graph traversal (connected via entity links)
-
-    Returns:
-        List of related observations with their tags for LLM tag routing
-    """
-    # Use recall to find related observations
-    # NO tags parameter - we want ALL observations regardless of scope
-    # Use low max_tokens since we only need observations, not memories
-    recall_result = await memory_engine.recall_async(
-        bank_id=bank_id,
-        query=query,
-        max_tokens=5000,  # Token budget for observations
-        fact_type=["observation"],  # Only retrieve observations
-        request_context=request_context,
-        _quiet=True,  # Suppress logging
-        # NO tags parameter - intentionally get ALL observations
-    )
-
-    # If no observations returned, return empty list
-    # When fact_type=["observation"], results come back in `results` field
-    if not recall_result.results:
-        return []
-
-    # Trust recall's relevance filtering - fetch full data for each observation
-    results = []
-    for obs in recall_result.results:
-        # Fetch full observation data from DB to get history, source_memory_ids, tags
-        row = await conn.fetchrow(
-            f"""
-            SELECT id, text, proof_count, history, tags, source_memory_ids, created_at, updated_at
-            FROM {fq_table("memory_units")}
-            WHERE id = $1 AND bank_id = $2 AND fact_type = 'observation'
-            """,
-            uuid.UUID(obs.id),
-            bank_id,
-        )
-
-        if row:
-            history = row["history"]
-            if isinstance(history, str):
-                history = json.loads(history)
-            elif history is None:
-                history = []
-
-            results.append(
-                {
-                    "id": row["id"],
-                    "text": row["text"],
-                    "proof_count": row["proof_count"] or 1,
-                    "history": history,
-                    "tags": row["tags"] or [],  # Include tags for LLM tag routing
-                    "source_memory_ids": row["source_memory_ids"] or [],
-                    "similarity": 1.0,  # Retrieved via recall so assumed relevant
-                }
-            )
-
-    return results
-
-
-async def _consolidate_with_llm(
-    memory_engine: "MemoryEngine",
-    fact_text: str,
-    observations: list[dict[str, Any]],
-    mission: str,
-) -> list[dict[str, Any]]:
-    """
-    Single LLM call to extract durable knowledge and decide on consolidation actions.
-
-    This handles ALL cases:
-    - No related observations: extracts durable knowledge, returns create action
-    - Related observations exist: compares and returns update/create actions
-    - Purely ephemeral fact: returns empty array
-
-    Note: Tags are NOT handled by the LLM. They are determined algorithmically:
-    - CREATE: observation inherits source fact's tags
-    - UPDATE: observation merges source fact's tags with existing tags
-
-    Returns:
-        List of actions, each being:
-        - {"action": "update", "learning_id": "uuid", "text": "...", "reason": "..."}
-        - {"action": "create", "text": "...", "reason": "..."}
-        - [] if fact is purely ephemeral (no durable knowledge)
-    """
-    # Format observations WITH their tags (or "None" if empty)
-    if observations:
-        observations_text = "\n".join(
-            f'- ID: {obs["id"]}, Tags: {json.dumps(obs["tags"])}, Text: "{obs["text"]}" (proof_count: {obs["proof_count"]})'
-            for obs in observations
-        )
-    else:
-        observations_text = "None (this is a new topic - create if fact contains durable knowledge)"
-
-    # Only include mission section if mission is set and not the default
-    mission_section = ""
-    if mission and mission != "General memory consolidation":
-        mission_section = f"""
-MISSION CONTEXT: {mission}
-
-Focus on DURABLE knowledge that serves this mission, not ephemeral state.
-"""
-
-    user_prompt = CONSOLIDATION_USER_PROMPT.format(
-        mission_section=mission_section,
-        fact_text=fact_text,
-        observations_text=observations_text,
-    )
-
-    messages = [
-        {"role": "system", "content": CONSOLIDATION_SYSTEM_PROMPT},
-        {"role": "user", "content": user_prompt},
-    ]
-
-    try:
-        result = await memory_engine._consolidation_llm_config.call(
-            messages=messages,
-            skip_validation=True,  # Raw JSON response
-            scope="consolidation",
-        )
-        # Parse JSON response - should be an array
-        if isinstance(result, str):
-            result = json.loads(result)
-        # Ensure result is a list
-        if isinstance(result, list):
-            return result
-        # Handle legacy single-action format for backward compatibility
-        if isinstance(result, dict):
-            if result.get("related_ids") and result.get("consolidated_text"):
-                # Convert old format to new format
-                return [
-                    {
-                        "action": "update",
-                        "learning_id": result["related_ids"][0],
-                        "text": result["consolidated_text"],
-                        "reason": result.get("reason", ""),
-                    }
-                ]
-            return []
-        return []
-    except Exception as e:
-        logger.warning(f"Error in consolidation LLM call: {e}")
-        return []
-
-
-async def _create_observation_directly(
-    conn: "Connection",
-    memory_engine: "MemoryEngine",
-    bank_id: str,
-    source_memory_id: uuid.UUID,
-    observation_text: str,
-    tags: list[str] | None = None,
-    event_date: datetime | None = None,
-    occurred_start: datetime | None = None,
-    occurred_end: datetime | None = None,
-    mentioned_at: datetime | None = None,
-    perf: ConsolidationPerfLog | None = None,
-) -> dict[str, Any]:
-    """
-    Create an observation directly with pre-processed text (no LLM call).
-
-    Used when the classify LLM has already provided the learning text.
-    This avoids the redundant second LLM call.
-    """
-    # Generate embedding for the observation (convert to string for pgvector)
-    t0 = time.time()
-    embeddings = await embedding_utils.generate_embeddings_batch(memory_engine.embeddings, [observation_text])
-    embedding_str = str(embeddings[0]) if embeddings else None
-    if perf:
-        perf.record_timing("embedding", time.time() - t0)
-
-    # Create the observation as a memory_unit
-    now = datetime.now(timezone.utc)
-    obs_event_date = event_date or now
-    obs_occurred_start = occurred_start or now
-    obs_occurred_end = occurred_end or now
-    obs_mentioned_at = mentioned_at or now
-    obs_tags = tags or []
-
-    t0 = time.time()
-    observation_id = uuid.uuid4()
-    row = await conn.fetchrow(
-        f"""
-        INSERT INTO {fq_table("memory_units")} (
-            id, bank_id, text, fact_type, embedding, proof_count, source_memory_ids, history,
-            tags, event_date, occurred_start, occurred_end, mentioned_at
-        )
-        VALUES ($1, $2, $3, 'observation', $4::vector, 1, $5, '[]'::jsonb, $6, $7, $8, $9, $10)
-        RETURNING id
-        """,
-        observation_id,
-        bank_id,
-        observation_text,
-        embedding_str,
-        [source_memory_id],
-        obs_tags,
-        obs_event_date,
-        obs_occurred_start,
-        obs_occurred_end,
-        obs_mentioned_at,
-    )
-
-    # Create links between memory and observation (includes entity links, memory_links)
-    await _create_memory_links(conn, source_memory_id, observation_id)
-    if perf:
-        perf.record_timing("db_write", time.time() - t0)
-
-    logger.debug(f"Created observation {observation_id} from memory {source_memory_id} (tags: {obs_tags})")
-
-    return {"action": "created", "observation_id": str(row["id"]), "tags": obs_tags}
@@ -1,69 +0,0 @@
-"""Prompts for the consolidation engine."""
-
-CONSOLIDATION_SYSTEM_PROMPT = """You are a memory consolidation system. Your job is to convert facts into durable knowledge (observations) and merge with existing knowledge when appropriate.
-
-You must output ONLY valid JSON with no markdown formatting, no code blocks, and no additional text.
-
-## EXTRACT DURABLE KNOWLEDGE, NOT EPHEMERAL STATE
-Facts often describe events or actions. Extract the DURABLE KNOWLEDGE implied by the fact, not the transient state.
-
-Examples of extracting durable knowledge:
- "User moved to Room 203" -> "Room 203 exists" (location exists, not where user is now)
- "User visited Acme Corp at Room 105" -> "Acme Corp is located in Room 105"
- "User took the elevator to floor 3" -> "Floor 3 is accessible by elevator"
- "User met Sarah at the lobby" -> "Sarah can be found at the lobby"
-
-DO NOT track current user position/state as knowledge - that changes constantly.
-DO track permanent facts learned from the user's actions.
-
-## PRESERVE SPECIFIC DETAILS
-Keep names, locations, numbers, and other specifics. Do NOT:
- Abstract into general principles
- Generate business insights
- Make knowledge generic
-
-GOOD examples:
- Fact: "John likes pizza" -> "John likes pizza"
- Fact: "Alice works at Google" -> "Alice works at Google"
-
-BAD examples:
- "John likes pizza" -> "Understanding dietary preferences helps..." (TOO ABSTRACT)
- "User is at Room 203" -> "User is currently at Room 203" (EPHEMERAL STATE)
-
-## MERGE RULES (when comparing to existing observations):
-1. REDUNDANT: Same information worded differently → update existing
-2. CONTRADICTION: Opposite information about same topic → update with history (e.g., "used to X, now Y")
-3. UPDATE: New state replacing old state → update with history
-
-## CRITICAL RULES:
- NEVER merge facts about DIFFERENT people
- NEVER merge unrelated topics (food preferences vs work vs hobbies)
- When merging contradictions, capture the CHANGE (before → after)
- Keep observations focused on ONE specific topic per person
- The "text" field MUST contain durable knowledge, not ephemeral state
- Do NOT include "tags" in output - tags are handled automatically"""
-
-CONSOLIDATION_USER_PROMPT = """Analyze this new fact and consolidate into knowledge.
-{mission_section}
-NEW FACT: {fact_text}
-
-EXISTING OBSERVATIONS:
-{observations_text}
-
-Instructions:
-1. First, extract the DURABLE KNOWLEDGE from the fact (not ephemeral state like "user is at X")
-2. Then compare with existing observations:
-   - If an observation covers the same topic: UPDATE it with the new knowledge
-   - If no observation covers the topic: CREATE a new one
-
-Output JSON array of actions (ALWAYS an array, even for single action):
-[
-  {{"action": "update", "learning_id": "uuid", "text": "updated durable knowledge", "reason": "..."}},
-  {{"action": "create", "text": "new durable knowledge", "reason": "..."}}
-]
-
-If NO consolidation is needed (fact is purely ephemeral with no durable knowledge):
-[]
-
-If no observations exist and fact contains durable knowledge:
-[{{"action": "create", "text": "durable knowledge text", "reason": "new topic"}}]"""
@@ -15,24 +15,18 @@ from concurrent.futures import ThreadPoolExecutor
 import httpx

 from ..config import (
-    DEFAULT_LITELLM_API_BASE,
    DEFAULT_RERANKER_COHERE_MODEL,
    DEFAULT_RERANKER_FLASHRANK_CACHE_DIR,
    DEFAULT_RERANKER_FLASHRANK_MODEL,
-    DEFAULT_RERANKER_LITELLM_MODEL,
    DEFAULT_RERANKER_LOCAL_MAX_CONCURRENT,
    DEFAULT_RERANKER_LOCAL_MODEL,
    DEFAULT_RERANKER_PROVIDER,
    DEFAULT_RERANKER_TEI_BATCH_SIZE,
    DEFAULT_RERANKER_TEI_MAX_CONCURRENT,
    ENV_COHERE_API_KEY,
-    ENV_LITELLM_API_BASE,
-    ENV_LITELLM_API_KEY,
-    ENV_RERANKER_COHERE_BASE_URL,
    ENV_RERANKER_COHERE_MODEL,
    ENV_RERANKER_FLASHRANK_CACHE_DIR,
    ENV_RERANKER_FLASHRANK_MODEL,
-    ENV_RERANKER_LITELLM_MODEL,
    ENV_RERANKER_LOCAL_MAX_CONCURRENT,
    ENV_RERANKER_LOCAL_MODEL,
    ENV_RERANKER_PROVIDER,
@@ -130,28 +124,13 @@ class LocalSTCrossEncoder(CrossEncoderModel):
                "Install it with: pip install sentence-transformers"
            )

+        # Note: We use CPU even when GPU/MPS is available because:
+        # 1. The reranker model (MiniLM) is tiny (~22M params)
+        # 2. Batch sizes are small (~100-200 pairs)
+        # 3. Data transfer overhead to GPU outweighs compute benefit
+        # 4. CPU inference is actually faster for this workload
        logger.info(f"Reranker: initializing local provider with model {self.model_name}")
-
-        # Determine device based on hardware availability.
-        # We always set low_cpu_mem_usage=False to prevent lazy loading (meta tensors)
-        # which can cause issues when accelerate is installed but no GPU is available.
-        # Note: We do NOT use device_map because CrossEncoder internally calls .to(device)
-        # after loading, which conflicts with accelerate's device_map handling.
-        import torch
-
-        # Check for GPU (CUDA) or Apple Silicon (MPS)
-        has_gpu = torch.cuda.is_available() or (hasattr(torch.backends, "mps") and torch.backends.mps.is_available())
-
-        if has_gpu:
-            device = None  # Let sentence-transformers auto-detect GPU/MPS
-        else:
-            device = "cpu"
-
-        self._model = CrossEncoder(
-            self.model_name,
-            device=device,
-            model_kwargs={"low_cpu_mem_usage": False},
-        )
+        self._model = CrossEncoder(self.model_name)

        # Initialize shared executor (limited workers naturally limits concurrency)
        if LocalSTCrossEncoder._executor is None:
@@ -163,101 +142,11 @@ class LocalSTCrossEncoder(CrossEncoderModel):
        else:
            logger.info("Reranker: local provider initialized (using existing executor)")

-    def _is_xpc_error(self, error: Exception) -> bool:
-        """
-        Check if an error is an XPC connection error (macOS daemon issue).
-
-        On macOS, long-running daemons can lose XPC connections to system services
-        when the process is idle for extended periods.
-        """
-        error_str = str(error).lower()
-        return "xpc_error_connection_invalid" in error_str or "xpc error" in error_str
-
-    def _reinitialize_model_sync(self) -> None:
-        """
-        Clear and reinitialize the cross-encoder model synchronously.
-
-        This is used to recover from XPC errors on macOS where the
-        PyTorch/MPS backend loses its connection to system services.
-        """
-        logger.warning(f"Reinitializing reranker model {self.model_name} due to backend error")
-
-        # Clear existing model
-        self._model = None
-
-        # Force garbage collection to free resources
-        import gc
-
-        import torch
-
-        gc.collect()
-
-        # If using CUDA/MPS, clear the cache
-        if torch.cuda.is_available():
-            torch.cuda.empty_cache()
-        elif hasattr(torch.backends, "mps") and torch.backends.mps.is_available():
-            try:
-                torch.mps.empty_cache()
-            except AttributeError:
-                pass  # Method might not exist in all PyTorch versions
-
-        # Reinitialize the model
-        try:
-            from sentence_transformers import CrossEncoder
-        except ImportError:
-            raise ImportError(
-                "sentence-transformers is required for LocalSTCrossEncoder. "
-                "Install it with: pip install sentence-transformers"
-            )
-
-        # Determine device based on hardware availability
-        has_gpu = torch.cuda.is_available() or (hasattr(torch.backends, "mps") and torch.backends.mps.is_available())
-
-        if has_gpu:
-            device = None  # Let sentence-transformers auto-detect GPU/MPS
-        else:
-            device = "cpu"
-
-        self._model = CrossEncoder(
-            self.model_name,
-            device=device,
-            model_kwargs={"low_cpu_mem_usage": False},
-        )
-
-        logger.info("Reranker: local provider reinitialized successfully")
-
-    def _predict_with_recovery(self, pairs: list[tuple[str, str]]) -> list[float]:
-        """
-        Predict with automatic recovery from XPC errors.
-
-        This runs synchronously in the thread pool.
-        """
-        max_retries = 1
-        for attempt in range(max_retries + 1):
-            try:
-                scores = self._model.predict(pairs, show_progress_bar=False)
-                return scores.tolist() if hasattr(scores, "tolist") else list(scores)
-            except Exception as e:
-                # Check if this is an XPC error (macOS daemon issue)
-                if self._is_xpc_error(e) and attempt < max_retries:
-                    logger.warning(f"XPC error detected in reranker (attempt {attempt + 1}): {e}")
-                    try:
-                        self._reinitialize_model_sync()
-                        logger.info("Reranker reinitialized successfully, retrying prediction")
-                        continue
-                    except Exception as reinit_error:
-                        logger.error(f"Failed to reinitialize reranker: {reinit_error}")
-                        raise Exception(f"Failed to recover from XPC error: {str(e)}")
-                else:
-                    # Not an XPC error or out of retries
-                    raise
-
    async def predict(self, pairs: list[tuple[str, str]]) -> list[float]:
        """
        Score query-document pairs for relevance.

        Uses a dedicated thread pool with limited workers to prevent CPU thrashing.
-        Automatically recovers from XPC errors on macOS by reinitializing the model.

        Args:
            pairs: List of (query, document) tuples to score
@@ -270,11 +159,11 @@ class LocalSTCrossEncoder(CrossEncoderModel):

        # Use dedicated executor - limited workers naturally limits concurrency
        loop = asyncio.get_event_loop()
-        return await loop.run_in_executor(
+        scores = await loop.run_in_executor(
            LocalSTCrossEncoder._executor,
-            self._predict_with_recovery,
-            pairs,
+            lambda: self._model.predict(pairs, show_progress_bar=False),
        )
+        return scores.tolist() if hasattr(scores, "tolist") else list(scores)


 class RemoteTEICrossEncoder(CrossEncoderModel):
@@ -503,7 +392,6 @@ class CohereCrossEncoder(CrossEncoderModel):
        self,
        api_key: str,
        model: str = DEFAULT_RERANKER_COHERE_MODEL,
-        base_url: str | None = None,
        timeout: float = 60.0,
    ):
        """
@@ -512,12 +400,10 @@ class CohereCrossEncoder(CrossEncoderModel):
        Args:
            api_key: Cohere API key
            model: Cohere rerank model name (default: rerank-english-v3.0)
-            base_url: Custom base URL for Cohere-compatible API (e.g., Azure-hosted endpoint)
            timeout: Request timeout in seconds (default: 60.0)
        """
        self.api_key = api_key
        self.model = model
-        self.base_url = base_url
        self.timeout = timeout
        self._client = None

@@ -535,14 +421,8 @@ class CohereCrossEncoder(CrossEncoderModel):
        except ImportError:
            raise ImportError("cohere is required for CohereCrossEncoder. Install it with: pip install cohere")

-        base_url_msg = f" at {self.base_url}" if self.base_url else ""
-        logger.info(f"Reranker: initializing Cohere provider with model {self.model}{base_url_msg}")
-
-        # Build client kwargs, only including base_url if set (for Azure or custom endpoints)
-        client_kwargs = {"api_key": self.api_key, "timeout": self.timeout}
-        if self.base_url:
-            client_kwargs["base_url"] = self.base_url
-        self._client = cohere.Client(**client_kwargs)
+        logger.info(f"Reranker: initializing Cohere provider with model {self.model}")
+        self._client = cohere.Client(api_key=self.api_key, timeout=self.timeout)
        logger.info("Reranker: Cohere provider initialized")

    async def predict(self, pairs: list[tuple[str, str]]) -> list[float]:
@@ -761,116 +641,6 @@ class FlashRankCrossEncoder(CrossEncoderModel):
        return await loop.run_in_executor(FlashRankCrossEncoder._executor, self._predict_sync, pairs)


-class LiteLLMCrossEncoder(CrossEncoderModel):
-    """
-    LiteLLM cross-encoder implementation using LiteLLM proxy's /rerank endpoint.
-
-    LiteLLM provides a unified interface for multiple reranking providers via
-    the Cohere-compatible /rerank endpoint.
-    See: https://docs.litellm.ai/docs/rerank
-
-    Supported providers via LiteLLM:
-    - Cohere (rerank-english-v3.0, etc.) - prefix with cohere/
-    - Together AI - prefix with together_ai/
-    - Azure AI - prefix with azure_ai/
-    - Jina AI - prefix with jina_ai/
-    - AWS Bedrock - prefix with bedrock/
-    - Voyage AI - prefix with voyage/
-    """
-
-    def __init__(
-        self,
-        api_base: str = DEFAULT_LITELLM_API_BASE,
-        api_key: str | None = None,
-        model: str = DEFAULT_RERANKER_LITELLM_MODEL,
-        timeout: float = 60.0,
-    ):
-        """
-        Initialize LiteLLM cross-encoder client.
-
-        Args:
-            api_base: Base URL of the LiteLLM proxy (default: http://localhost:4000)
-            api_key: API key for the LiteLLM proxy (optional, depends on proxy config)
-            model: Reranking model name (default: cohere/rerank-english-v3.0)
-                   Use provider prefix (e.g., cohere/, together_ai/, voyage/)
-            timeout: Request timeout in seconds (default: 60.0)
-        """
-        self.api_base = api_base.rstrip("/")
-        self.api_key = api_key
-        self.model = model
-        self.timeout = timeout
-        self._async_client: httpx.AsyncClient | None = None
-
-    @property
-    def provider_name(self) -> str:
-        return "litellm"
-
-    async def initialize(self) -> None:
-        """Initialize the async HTTP client."""
-        if self._async_client is not None:
-            return
-
-        logger.info(f"Reranker: initializing LiteLLM provider at {self.api_base} with model {self.model}")
-
-        headers = {"Content-Type": "application/json"}
-        if self.api_key:
-            headers["Authorization"] = f"Bearer {self.api_key}"
-
-        self._async_client = httpx.AsyncClient(timeout=self.timeout, headers=headers)
-        logger.info("Reranker: LiteLLM provider initialized")
-
-    async def predict(self, pairs: list[tuple[str, str]]) -> list[float]:
-        """
-        Score query-document pairs using the LiteLLM proxy's /rerank endpoint.
-
-        Args:
-            pairs: List of (query, document) tuples to score
-
-        Returns:
-            List of relevance scores
-        """
-        if self._async_client is None:
-            raise RuntimeError("Reranker not initialized. Call initialize() first.")
-
-        if not pairs:
-            return []
-
-        # Group pairs by query (LiteLLM rerank expects one query with multiple documents)
-        query_groups: dict[str, list[tuple[int, str]]] = {}
-        for idx, (query, text) in enumerate(pairs):
-            if query not in query_groups:
-                query_groups[query] = []
-            query_groups[query].append((idx, text))
-
-        all_scores = [0.0] * len(pairs)
-
-        for query, indexed_texts in query_groups.items():
-            texts = [text for _, text in indexed_texts]
-            indices = [idx for idx, _ in indexed_texts]
-
-            # LiteLLM /rerank follows Cohere API format
-            response = await self._async_client.post(
-                f"{self.api_base}/rerank",
-                json={
-                    "model": self.model,
-                    "query": query,
-                    "documents": texts,
-                    "top_n": len(texts),  # Return all scores
-                },
-            )
-            response.raise_for_status()
-            result = response.json()
-
-            # Map scores back to original positions
-            # Response format: {"results": [{"index": 0, "relevance_score": 0.9}, ...]}
-            for item in result.get("results", []):
-                original_idx = item["index"]
-                score = item.get("relevance_score", item.get("score", 0.0))
-                all_scores[indices[original_idx]] = score
-
-        return all_scores
-
-
 def create_cross_encoder_from_env() -> CrossEncoderModel:
    """
    Create a CrossEncoderModel instance based on environment variables.
@@ -901,20 +671,14 @@ def create_cross_encoder_from_env() -> CrossEncoderModel:
        if not api_key:
            raise ValueError(f"{ENV_COHERE_API_KEY} is required when {ENV_RERANKER_PROVIDER} is 'cohere'")
        model = os.environ.get(ENV_RERANKER_COHERE_MODEL, DEFAULT_RERANKER_COHERE_MODEL)
-        base_url = os.environ.get(ENV_RERANKER_COHERE_BASE_URL) or None
-        return CohereCrossEncoder(api_key=api_key, model=model, base_url=base_url)
+        return CohereCrossEncoder(api_key=api_key, model=model)
    elif provider == "flashrank":
        model = os.environ.get(ENV_RERANKER_FLASHRANK_MODEL, DEFAULT_RERANKER_FLASHRANK_MODEL)
        cache_dir = os.environ.get(ENV_RERANKER_FLASHRANK_CACHE_DIR, DEFAULT_RERANKER_FLASHRANK_CACHE_DIR)
        return FlashRankCrossEncoder(model_name=model, cache_dir=cache_dir)
-    elif provider == "litellm":
-        api_base = os.environ.get(ENV_LITELLM_API_BASE, DEFAULT_LITELLM_API_BASE)
-        api_key = os.environ.get(ENV_LITELLM_API_KEY)
-        model = os.environ.get(ENV_RERANKER_LITELLM_MODEL, DEFAULT_RERANKER_LITELLM_MODEL)
-        return LiteLLMCrossEncoder(api_base=api_base, api_key=api_key, model=model)
    elif provider == "rrf":
        return RRFPassthroughCrossEncoder()
    else:
        raise ValueError(
-            f"Unknown reranker provider: {provider}. Supported: 'local', 'tei', 'cohere', 'flashrank', 'litellm', 'rrf'"
+            f"Unknown reranker provider: {provider}. Supported: 'local', 'tei', 'cohere', 'flashrank', 'rrf'"
        )
@@ -1,5 +0,0 @@
-"""Directives module for hard rules injected into prompts."""
-
-from .models import Directive
-
-__all__ = ["Directive"]
@@ -1,37 +0,0 @@
-"""Pydantic models for directives."""
-
-from datetime import datetime, timezone
-from uuid import UUID
-
-from pydantic import BaseModel, Field
-
-
-class Directive(BaseModel):
-    """A directive is a hard rule injected into prompts.
-
-    Directives are user-defined rules that guide agent behavior. Unlike mental models
-    which are automatically consolidated from memories, directives are explicit
-    instructions that are always included in relevant prompts.
-
-    Examples:
-    - "Always respond in formal English"
-    - "Never share personal data with third parties"
-    - "Prefer conservative investment recommendations"
-    """
-
-    id: UUID = Field(description="Unique identifier")
-    bank_id: str = Field(description="Bank this directive belongs to")
-    name: str = Field(description="Human-readable name")
-    content: str = Field(description="The directive text to inject into prompts")
-    priority: int = Field(default=0, description="Higher priority directives are injected first")
-    is_active: bool = Field(default=True, description="Whether this directive is currently active")
-    tags: list[str] = Field(default_factory=list, description="Tags for filtering")
-    created_at: datetime = Field(
-        default_factory=lambda: datetime.now(timezone.utc), description="When this directive was created"
-    )
-    updated_at: datetime = Field(
-        default_factory=lambda: datetime.now(timezone.utc), description="When this directive was last updated"
-    )
-
-    class Config:
-        from_attributes = True
@@ -17,23 +17,16 @@ import httpx

 from ..config import (
    DEFAULT_EMBEDDINGS_COHERE_MODEL,
-    DEFAULT_EMBEDDINGS_LITELLM_MODEL,
    DEFAULT_EMBEDDINGS_LOCAL_MODEL,
    DEFAULT_EMBEDDINGS_OPENAI_MODEL,
    DEFAULT_EMBEDDINGS_PROVIDER,
-    DEFAULT_LITELLM_API_BASE,
    ENV_COHERE_API_KEY,
-    ENV_EMBEDDINGS_COHERE_BASE_URL,
    ENV_EMBEDDINGS_COHERE_MODEL,
-    ENV_EMBEDDINGS_LITELLM_MODEL,
    ENV_EMBEDDINGS_LOCAL_MODEL,
    ENV_EMBEDDINGS_OPENAI_API_KEY,
-    ENV_EMBEDDINGS_OPENAI_BASE_URL,
    ENV_EMBEDDINGS_OPENAI_MODEL,
    ENV_EMBEDDINGS_PROVIDER,
    ENV_EMBEDDINGS_TEI_URL,
-    ENV_LITELLM_API_BASE,
-    ENV_LITELLM_API_KEY,
    ENV_LLM_API_KEY,
 )

@@ -128,98 +121,20 @@ class LocalSTEmbeddings(Embeddings):
            )

        logger.info(f"Embeddings: initializing local provider with model {self.model_name}")
-
-        # Determine device based on hardware availability.
-        # We always set low_cpu_mem_usage=False to prevent lazy loading (meta tensors)
-        # which can cause issues when accelerate is installed but no GPU is available.
-        import torch
-
-        # Check for GPU (CUDA) or Apple Silicon (MPS)
-        has_gpu = torch.cuda.is_available() or (hasattr(torch.backends, "mps") and torch.backends.mps.is_available())
-
-        if has_gpu:
-            device = None  # Let sentence-transformers auto-detect GPU/MPS
-        else:
-            device = "cpu"
-
+        # Disable lazy loading (meta tensors) which causes issues with newer transformers/accelerate
+        # Setting low_cpu_mem_usage=False and device_map=None ensures tensors are fully materialized
        self._model = SentenceTransformer(
            self.model_name,
-            device=device,
-            model_kwargs={"low_cpu_mem_usage": False},
+            model_kwargs={"low_cpu_mem_usage": False, "device_map": None},
        )

        self._dimension = self._model.get_sentence_embedding_dimension()
        logger.info(f"Embeddings: local provider initialized (dim: {self._dimension})")

-    def _is_xpc_error(self, error: Exception) -> bool:
-        """
-        Check if an error is an XPC connection error (macOS daemon issue).
-
-        On macOS, long-running daemons can lose XPC connections to system services
-        when the process is idle for extended periods.
-        """
-        error_str = str(error).lower()
-        return "xpc_error_connection_invalid" in error_str or "xpc error" in error_str
-
-    def _reinitialize_model_sync(self) -> None:
-        """
-        Clear and reinitialize the embedding model synchronously.
-
-        This is used to recover from XPC errors on macOS where the
-        PyTorch/MPS backend loses its connection to system services.
-        """
-        logger.warning(f"Reinitializing embedding model {self.model_name} due to backend error")
-
-        # Clear existing model
-        self._model = None
-
-        # Force garbage collection to free resources
-        import gc
-
-        import torch
-
-        gc.collect()
-
-        # If using CUDA/MPS, clear the cache
-        if torch.cuda.is_available():
-            torch.cuda.empty_cache()
-        elif hasattr(torch.backends, "mps") and torch.backends.mps.is_available():
-            try:
-                torch.mps.empty_cache()
-            except AttributeError:
-                pass  # Method might not exist in all PyTorch versions
-
-        # Reinitialize the model (inline version of initialize() but synchronous)
-        try:
-            from sentence_transformers import SentenceTransformer
-        except ImportError:
-            raise ImportError(
-                "sentence-transformers is required for LocalSTEmbeddings. "
-                "Install it with: pip install sentence-transformers"
-            )
-
-        # Determine device based on hardware availability
-        has_gpu = torch.cuda.is_available() or (hasattr(torch.backends, "mps") and torch.backends.mps.is_available())
-
-        if has_gpu:
-            device = None  # Let sentence-transformers auto-detect GPU/MPS
-        else:
-            device = "cpu"
-
-        self._model = SentenceTransformer(
-            self.model_name,
-            device=device,
-            model_kwargs={"low_cpu_mem_usage": False},
-        )
-
-        logger.info("Embeddings: local provider reinitialized successfully")
-
    def encode(self, texts: list[str]) -> list[list[float]]:
        """
        Generate embeddings for a list of texts.

-        Automatically recovers from XPC errors on macOS by reinitializing the model.
-
        Args:
            texts: List of text strings to encode

@@ -228,27 +143,8 @@ class LocalSTEmbeddings(Embeddings):
        """
        if self._model is None:
            raise RuntimeError("Embeddings not initialized. Call initialize() first.")
-
-        # Try encoding with automatic recovery from XPC errors
-        max_retries = 1
-        for attempt in range(max_retries + 1):
-            try:
-                embeddings = self._model.encode(texts, convert_to_numpy=True, show_progress_bar=False)
-                return [emb.tolist() for emb in embeddings]
-            except Exception as e:
-                # Check if this is an XPC error (macOS daemon issue)
-                if self._is_xpc_error(e) and attempt < max_retries:
-                    logger.warning(f"XPC error detected in embedding generation (attempt {attempt + 1}): {e}")
-                    try:
-                        self._reinitialize_model_sync()
-                        logger.info("Model reinitialized successfully, retrying embedding generation")
-                        continue
-                    except Exception as reinit_error:
-                        logger.error(f"Failed to reinitialize model: {reinit_error}")
-                        raise Exception(f"Failed to recover from XPC error: {str(e)}")
-                else:
-                    # Not an XPC error or out of retries
-                    raise
+        embeddings = self._model.encode(texts, convert_to_numpy=True, show_progress_bar=False)
+        return [emb.tolist() for emb in embeddings]


 class RemoteTEIEmbeddings(Embeddings):
@@ -426,7 +322,6 @@ class OpenAIEmbeddings(Embeddings):
        self,
        api_key: str,
        model: str = DEFAULT_EMBEDDINGS_OPENAI_MODEL,
-        base_url: str | None = None,
        batch_size: int = 100,
        max_retries: int = 3,
    ):
@@ -436,13 +331,11 @@ class OpenAIEmbeddings(Embeddings):
        Args:
            api_key: OpenAI API key
            model: OpenAI embedding model name (default: text-embedding-3-small)
-            base_url: Custom base URL for OpenAI-compatible API (e.g., Azure OpenAI endpoint)
            batch_size: Maximum batch size for embedding requests (default: 100)
            max_retries: Maximum number of retries for failed requests (default: 3)
        """
        self.api_key = api_key
        self.model = model
-        self.base_url = base_url
        self.batch_size = batch_size
        self.max_retries = max_retries
        self._client = None
@@ -468,14 +361,8 @@ class OpenAIEmbeddings(Embeddings):
        except ImportError:
            raise ImportError("openai is required for OpenAIEmbeddings. Install it with: pip install openai")

-        base_url_msg = f" at {self.base_url}" if self.base_url else ""
-        logger.info(f"Embeddings: initializing OpenAI provider with model {self.model}{base_url_msg}")
-
-        # Build client kwargs, only including base_url if set (for Azure or custom endpoints)
-        client_kwargs = {"api_key": self.api_key, "max_retries": self.max_retries}
-        if self.base_url:
-            client_kwargs["base_url"] = self.base_url
-        self._client = OpenAI(**client_kwargs)
+        logger.info(f"Embeddings: initializing OpenAI provider with model {self.model}")
+        self._client = OpenAI(api_key=self.api_key, max_retries=self.max_retries)

        # Try to get dimension from known models, otherwise do a test embedding
        if self.model in self.MODEL_DIMENSIONS:
@@ -548,7 +435,6 @@ class CohereEmbeddings(Embeddings):
        self,
        api_key: str,
        model: str = DEFAULT_EMBEDDINGS_COHERE_MODEL,
-        base_url: str | None = None,
        batch_size: int = 96,
        timeout: float = 60.0,
        input_type: str = "search_document",
@@ -559,7 +445,6 @@ class CohereEmbeddings(Embeddings):
        Args:
            api_key: Cohere API key
            model: Cohere embedding model name (default: embed-english-v3.0)
-            base_url: Custom base URL for Cohere-compatible API (e.g., Azure-hosted endpoint)
            batch_size: Maximum batch size for embedding requests (default: 96, Cohere's limit)
            timeout: Request timeout in seconds (default: 60.0)
            input_type: Input type for embeddings (default: search_document).
@@ -567,7 +452,6 @@ class CohereEmbeddings(Embeddings):
        """
        self.api_key = api_key
        self.model = model
-        self.base_url = base_url
        self.batch_size = batch_size
        self.timeout = timeout
        self.input_type = input_type
@@ -594,14 +478,8 @@ class CohereEmbeddings(Embeddings):
        except ImportError:
            raise ImportError("cohere is required for CohereEmbeddings. Install it with: pip install cohere")

-        base_url_msg = f" at {self.base_url}" if self.base_url else ""
-        logger.info(f"Embeddings: initializing Cohere provider with model {self.model}{base_url_msg}")
-
-        # Build client kwargs, only including base_url if set (for Azure or custom endpoints)
-        client_kwargs = {"api_key": self.api_key, "timeout": self.timeout}
-        if self.base_url:
-            client_kwargs["base_url"] = self.base_url
-        self._client = cohere.Client(**client_kwargs)
+        logger.info(f"Embeddings: initializing Cohere provider with model {self.model}")
+        self._client = cohere.Client(api_key=self.api_key, timeout=self.timeout)

        # Try to get dimension from known models, otherwise do a test embedding
        if self.model in self.MODEL_DIMENSIONS:
@@ -651,123 +529,6 @@ class CohereEmbeddings(Embeddings):
        return all_embeddings


-class LiteLLMEmbeddings(Embeddings):
-    """
-    LiteLLM embeddings implementation using LiteLLM proxy's /embeddings endpoint.
-
-    LiteLLM provides a unified interface for multiple embedding providers.
-    The proxy exposes an OpenAI-compatible /embeddings endpoint.
-    See: https://docs.litellm.ai/docs/embedding/supported_embedding
-
-    Supported providers via LiteLLM:
-    - OpenAI (text-embedding-3-small, text-embedding-ada-002, etc.)
-    - Cohere (embed-english-v3.0, etc.) - prefix with cohere/
-    - Vertex AI (textembedding-gecko, etc.) - prefix with vertex_ai/
-    - HuggingFace, Mistral, Voyage AI, etc.
-
-    The embedding dimension is auto-detected from the model at initialization.
-    """
-
-    def __init__(
-        self,
-        api_base: str = DEFAULT_LITELLM_API_BASE,
-        api_key: str | None = None,
-        model: str = DEFAULT_EMBEDDINGS_LITELLM_MODEL,
-        batch_size: int = 100,
-        timeout: float = 60.0,
-    ):
-        """
-        Initialize LiteLLM embeddings client.
-
-        Args:
-            api_base: Base URL of the LiteLLM proxy (default: http://localhost:4000)
-            api_key: API key for the LiteLLM proxy (optional, depends on proxy config)
-            model: Embedding model name (default: text-embedding-3-small)
-                   Use provider prefix for non-OpenAI models (e.g., cohere/embed-english-v3.0)
-            batch_size: Maximum batch size for embedding requests (default: 100)
-            timeout: Request timeout in seconds (default: 60.0)
-        """
-        self.api_base = api_base.rstrip("/")
-        self.api_key = api_key
-        self.model = model
-        self.batch_size = batch_size
-        self.timeout = timeout
-        self._client: httpx.Client | None = None
-        self._dimension: int | None = None
-
-    @property
-    def provider_name(self) -> str:
-        return "litellm"
-
-    @property
-    def dimension(self) -> int:
-        if self._dimension is None:
-            raise RuntimeError("Embeddings not initialized. Call initialize() first.")
-        return self._dimension
-
-    async def initialize(self) -> None:
-        """Initialize the HTTP client and detect embedding dimension."""
-        if self._client is not None:
-            return
-
-        logger.info(f"Embeddings: initializing LiteLLM provider at {self.api_base} with model {self.model}")
-
-        headers = {"Content-Type": "application/json"}
-        if self.api_key:
-            headers["Authorization"] = f"Bearer {self.api_key}"
-
-        self._client = httpx.Client(timeout=self.timeout, headers=headers)
-
-        # Do a test embedding to detect dimension
-        try:
-            response = self._client.post(
-                f"{self.api_base}/embeddings",
-                json={"model": self.model, "input": ["test"]},
-            )
-            response.raise_for_status()
-            result = response.json()
-            if result.get("data") and len(result["data"]) > 0:
-                self._dimension = len(result["data"][0]["embedding"])
-            logger.info(f"Embeddings: LiteLLM provider initialized (model: {self.model}, dim: {self._dimension})")
-        except httpx.HTTPError as e:
-            raise RuntimeError(f"Failed to connect to LiteLLM proxy at {self.api_base}: {e}")
-
-    def encode(self, texts: list[str]) -> list[list[float]]:
-        """
-        Generate embeddings using the LiteLLM proxy.
-
-        Args:
-            texts: List of text strings to encode
-
-        Returns:
-            List of embedding vectors
-        """
-        if self._client is None:
-            raise RuntimeError("Embeddings not initialized. Call initialize() first.")
-
-        if not texts:
-            return []
-
-        all_embeddings = []
-
-        # Process in batches
-        for i in range(0, len(texts), self.batch_size):
-            batch = texts[i : i + self.batch_size]
-
-            response = self._client.post(
-                f"{self.api_base}/embeddings",
-                json={"model": self.model, "input": batch},
-            )
-            response.raise_for_status()
-            result = response.json()
-
-            # Sort by index to ensure correct order
-            batch_embeddings = sorted(result["data"], key=lambda x: x["index"])
-            all_embeddings.extend([e["embedding"] for e in batch_embeddings])
-
-        return all_embeddings
-
-
 def create_embeddings_from_env() -> Embeddings:
    """
    Create an Embeddings instance based on environment variables.
@@ -797,21 +558,12 @@ def create_embeddings_from_env() -> Embeddings:
                f"when {ENV_EMBEDDINGS_PROVIDER} is 'openai'"
            )
        model = os.environ.get(ENV_EMBEDDINGS_OPENAI_MODEL, DEFAULT_EMBEDDINGS_OPENAI_MODEL)
-        base_url = os.environ.get(ENV_EMBEDDINGS_OPENAI_BASE_URL) or None
-        return OpenAIEmbeddings(api_key=api_key, model=model, base_url=base_url)
+        return OpenAIEmbeddings(api_key=api_key, model=model)
    elif provider == "cohere":
        api_key = os.environ.get(ENV_COHERE_API_KEY)
        if not api_key:
            raise ValueError(f"{ENV_COHERE_API_KEY} is required when {ENV_EMBEDDINGS_PROVIDER} is 'cohere'")
        model = os.environ.get(ENV_EMBEDDINGS_COHERE_MODEL, DEFAULT_EMBEDDINGS_COHERE_MODEL)
-        base_url = os.environ.get(ENV_EMBEDDINGS_COHERE_BASE_URL) or None
-        return CohereEmbeddings(api_key=api_key, model=model, base_url=base_url)
-    elif provider == "litellm":
-        api_base = os.environ.get(ENV_LITELLM_API_BASE, DEFAULT_LITELLM_API_BASE)
-        api_key = os.environ.get(ENV_LITELLM_API_KEY)
-        model = os.environ.get(ENV_EMBEDDINGS_LITELLM_MODEL, DEFAULT_EMBEDDINGS_LITELLM_MODEL)
-        return LiteLLMEmbeddings(api_base=api_base, api_key=api_key, model=model)
+        return CohereEmbeddings(api_key=api_key, model=model)
    else:
-        raise ValueError(
-            f"Unknown embeddings provider: {provider}. Supported: 'local', 'tei', 'openai', 'cohere', 'litellm'"
-        )
+        raise ValueError(f"Unknown embeddings provider: {provider}. Supported: 'local', 'tei', 'openai', 'cohere'")
@@ -160,14 +160,14 @@ class MemoryEngineInterface(ABC):
        request_context: "RequestContext",
    ) -> dict[str, Any]:
        """
-        Get bank profile including disposition and mission.
+        Get bank profile including disposition and background.

        Args:
            bank_id: The memory bank ID.
            request_context: Request context for authentication.

        Returns:
-            Bank profile dict with bank_id, name, disposition, and mission.
+            Bank profile dict.
        """
        ...

@@ -190,44 +190,25 @@ class MemoryEngineInterface(ABC):
        ...

    @abstractmethod
-    async def merge_bank_mission(
+    async def merge_bank_background(
        self,
        bank_id: str,
        new_info: str,
        *,
+        update_disposition: bool = True,
        request_context: "RequestContext",
    ) -> dict[str, Any]:
        """
-        Merge new mission information into bank profile.
+        Merge new background information into bank profile.

        Args:
            bank_id: The memory bank ID.
-            new_info: New mission information to merge.
+            new_info: New background information to merge.
+            update_disposition: Whether to infer disposition from background.
            request_context: Request context for authentication.

        Returns:
-            Updated mission info.
-        """
-        ...
-
-    @abstractmethod
-    async def set_bank_mission(
-        self,
-        bank_id: str,
-        mission: str,
-        *,
-        request_context: "RequestContext",
-    ) -> dict[str, Any]:
-        """
-        Set the bank's mission (replaces existing).
-
-        Args:
-            bank_id: The memory bank ID.
-            mission: The mission text.
-            request_context: Request context for authentication.
-
-        Returns:
-            Dict with bank_id and mission.
+            Updated background info.
        """
        ...

@@ -537,7 +518,7 @@ class MemoryEngineInterface(ABC):
        bank_id: str,
        *,
        request_context: "RequestContext",
-    ) -> dict[str, Any]:
+    ) -> list[dict[str, Any]]:
        """
        List async operations for a bank.

@@ -546,7 +527,7 @@ class MemoryEngineInterface(ABC):
            request_context: Request context for authentication.

        Returns:
-            Dict with 'total' (int) and 'operations' (list of operation dicts).
+            List of operation dicts with id, task_type, status, etc.
        """
        ...

@@ -580,16 +561,16 @@ class MemoryEngineInterface(ABC):
        bank_id: str,
        *,
        name: str | None = None,
-        mission: str | None = None,
+        background: str | None = None,
        request_context: "RequestContext",
    ) -> dict[str, Any]:
        """
-        Update bank name and/or mission.
+        Update bank name and/or background.

        Args:
            bank_id: The memory bank ID.
            name: New bank name (optional).
-            mission: New mission text (optional, replaces existing).
+            background: New background text (optional, replaces existing).
            request_context: Request context for authentication.

        Returns:
@@ -209,10 +209,10 @@ class LLMProvider:
            OutputTooLongError: If output exceeds token limits.
            Exception: Re-raises API errors after retries exhausted.
        """
-        semaphore_start = time.time()
+        queue_start_time = time.time()
        async with _global_llm_semaphore:
-            semaphore_wait_time = time.time() - semaphore_start
            start_time = time.time()
+            semaphore_wait_time = start_time - queue_start_time

            # Handle Mock provider (for testing)
            if self.provider == "mock":
@@ -318,44 +318,43 @@ class LLMProvider:

            last_exception = None

-            # Prepare response format ONCE before the retry loop
-            # (to avoid appending schema to messages on every retry)
-            if response_format is not None:
-                schema = None
-                if hasattr(response_format, "model_json_schema"):
-                    schema = response_format.model_json_schema()
-
-                if strict_schema and schema is not None:
-                    # Use OpenAI's strict JSON schema enforcement
-                    # This guarantees all required fields are returned
-                    call_params["response_format"] = {
-                        "type": "json_schema",
-                        "json_schema": {
-                            "name": "response",
-                            "strict": True,
-                            "schema": schema,
-                        },
-                    }
-                else:
-                    # Soft enforcement: add schema to prompt and use json_object mode
-                    if schema is not None:
-                        schema_msg = f"\n\nYou must respond with valid JSON matching this schema:\n{json.dumps(schema, indent=2)}"
-
-                        if call_params["messages"] and call_params["messages"][0].get("role") == "system":
-                            call_params["messages"][0]["content"] += schema_msg
-                        elif call_params["messages"]:
-                            call_params["messages"][0]["content"] = (
-                                schema_msg + "\n\n" + call_params["messages"][0]["content"]
-                            )
-                    if self.provider not in ("lmstudio", "ollama"):
-                        # LM Studio and Ollama don't support json_object response format reliably
-                        # We rely on the schema in the system message instead
-                        call_params["response_format"] = {"type": "json_object"}
-
            for attempt in range(max_retries + 1):
                try:
                    if response_format is not None:
+                        schema = None
+                        if hasattr(response_format, "model_json_schema"):
+                            schema = response_format.model_json_schema()
+
+                        if strict_schema and schema is not None:
+                            # Use OpenAI's strict JSON schema enforcement
+                            # This guarantees all required fields are returned
+                            call_params["response_format"] = {
+                                "type": "json_schema",
+                                "json_schema": {
+                                    "name": "response",
+                                    "strict": True,
+                                    "schema": schema,
+                                },
+                            }
+                        else:
+                            # Soft enforcement: add schema to prompt and use json_object mode
+                            if schema is not None:
+                                schema_msg = f"\n\nYou must respond with valid JSON matching this schema:\n{json.dumps(schema, indent=2)}"
+
+                                if call_params["messages"] and call_params["messages"][0].get("role") == "system":
+                                    call_params["messages"][0]["content"] += schema_msg
+                                elif call_params["messages"]:
+                                    call_params["messages"][0]["content"] = (
+                                        schema_msg + "\n\n" + call_params["messages"][0]["content"]
+                                    )
+                            if self.provider not in ("lmstudio", "ollama"):
+                                # LM Studio and Ollama don't support json_object response format reliably
+                                # We rely on the schema in the system message instead
+                                call_params["response_format"] = {"type": "json_object"}
+
+                        logger.debug(f"Sending request to {self.provider}/{self.model} (timeout={self.timeout})")
                        response = await self._client.chat.completions.create(**call_params)
+                        logger.debug(f"Received response from {self.provider}/{self.model}")

                        content = response.choices[0].message.content

@@ -468,11 +467,13 @@ class LLMProvider:

                except APIConnectionError as e:
                    last_exception = e
-                    status_code = getattr(e, "status_code", None) or getattr(
-                        getattr(e, "response", None), "status_code", None
-                    )
-                    logger.warning(f"APIConnectionError (HTTP {status_code}), attempt {attempt + 1}: {str(e)[:200]}")
                    if attempt < max_retries:
+                        status_code = getattr(e, "status_code", None) or getattr(
+                            getattr(e, "response", None), "status_code", None
+                        )
+                        logger.warning(
+                            f"Connection error, retrying... (attempt {attempt + 1}/{max_retries + 1}) - status_code={status_code}, message={e}"
+                        )
                        backoff = min(initial_backoff * (2**attempt), max_backoff)
                        await asyncio.sleep(backoff)
                        continue
@@ -486,45 +487,6 @@ class LLMProvider:
                        logger.error(f"Auth error (HTTP {e.status_code}), not retrying: {str(e)}")
                        raise

-                    # Handle tool_use_failed error - model outputted in tool call format
-                    # Convert to expected JSON format and continue
-                    if e.status_code == 400 and response_format is not None:
-                        try:
-                            error_body = e.body if hasattr(e, "body") else {}
-                            if isinstance(error_body, dict):
-                                error_info: dict[str, Any] = error_body.get("error") or {}
-                                if error_info.get("code") == "tool_use_failed":
-                                    failed_gen = error_info.get("failed_generation", "")
-                                    if failed_gen:
-                                        # Parse the tool call format and convert to actions format
-                                        tool_call = json.loads(failed_gen)
-                                        tool_name = tool_call.get("name", "")
-                                        tool_args = tool_call.get("arguments", {})
-                                        # Convert to actions format: {"actions": [{"tool": "name", ...args}]}
-                                        converted = {"actions": [{"tool": tool_name, **tool_args}]}
-                                        if skip_validation:
-                                            result = converted
-                                        else:
-                                            result = response_format.model_validate(converted)
-
-                                        # Record metrics for this successful recovery
-                                        duration = time.time() - start_time
-                                        metrics = get_metrics_collector()
-                                        metrics.record_llm_call(
-                                            provider=self.provider,
-                                            model=self.model,
-                                            scope=scope,
-                                            duration=duration,
-                                            input_tokens=0,
-                                            output_tokens=0,
-                                            success=True,
-                                        )
-                                        if return_usage:
-                                            return result, TokenUsage(input_tokens=0, output_tokens=0, total_tokens=0)
-                                        return result
-                        except (json.JSONDecodeError, KeyError, TypeError):
-                            pass  # Failed to parse tool_use_failed, continue with normal retry
-
                    last_exception = e
                    if attempt < max_retries:
                        backoff = min(initial_backoff * (2**attempt), max_backoff)
@@ -535,438 +497,14 @@ class LLMProvider:
                        logger.error(f"API error after {max_retries + 1} attempts: {str(e)}")
                        raise

-                except Exception:
+                except Exception as e:
+                    logger.error(f"Unexpected error during LLM call: {type(e).__name__}: {str(e)}")
                    raise

            if last_exception:
                raise last_exception
            raise RuntimeError("LLM call failed after all retries with no exception captured")

-    async def call_with_tools(
-        self,
-        messages: list[dict[str, Any]],
-        tools: list[dict[str, Any]],
-        max_completion_tokens: int | None = None,
-        temperature: float | None = None,
-        scope: str = "tools",
-        max_retries: int = 5,
-        initial_backoff: float = 1.0,
-        max_backoff: float = 30.0,
-        tool_choice: str | dict[str, Any] = "auto",
-    ) -> "LLMToolCallResult":
-        """
-        Make an LLM API call with tool/function calling support.
-
-        Args:
-            messages: List of message dicts. Can include tool results with role='tool'.
-            tools: List of tool definitions in OpenAI format.
-            max_completion_tokens: Maximum tokens in response.
-            temperature: Sampling temperature (0.0-2.0).
-            scope: Scope identifier for tracking.
-            max_retries: Maximum retry attempts.
-            initial_backoff: Initial backoff time in seconds.
-            max_backoff: Maximum backoff time in seconds.
-            tool_choice: How to choose tools - "auto", "none", "required", or {"type": "function", "function": {"name": "..."}}
-
-        Returns:
-            LLMToolCallResult with content and/or tool_calls.
-        """
-        from .response_models import LLMToolCall, LLMToolCallResult
-
-        async with _global_llm_semaphore:
-            start_time = time.time()
-
-            # Handle Mock provider
-            if self.provider == "mock":
-                return await self._call_with_tools_mock(messages, tools, scope)
-
-            # Handle Anthropic separately (uses different tool format)
-            if self.provider == "anthropic":
-                return await self._call_with_tools_anthropic(
-                    messages, tools, max_completion_tokens, max_retries, initial_backoff, max_backoff, start_time, scope
-                )
-
-            # Handle Gemini (convert to Gemini tool format)
-            if self.provider == "gemini":
-                return await self._call_with_tools_gemini(
-                    messages, tools, max_retries, initial_backoff, max_backoff, start_time, scope
-                )
-
-            # OpenAI-compatible providers (OpenAI, Groq, Ollama, LMStudio)
-            call_params: dict[str, Any] = {
-                "model": self.model,
-                "messages": messages,
-                "tools": tools,
-                "tool_choice": tool_choice,
-            }
-
-            if max_completion_tokens is not None:
-                call_params["max_completion_tokens"] = max_completion_tokens
-            if temperature is not None:
-                call_params["temperature"] = temperature
-
-            # Provider-specific parameters
-            if self.provider == "groq":
-                call_params["seed"] = DEFAULT_LLM_SEED
-
-            last_exception = None
-
-            for attempt in range(max_retries + 1):
-                try:
-                    response = await self._client.chat.completions.create(**call_params)
-
-                    message = response.choices[0].message
-                    finish_reason = response.choices[0].finish_reason
-
-                    # Extract tool calls if present
-                    tool_calls: list[LLMToolCall] = []
-                    if message.tool_calls:
-                        for tc in message.tool_calls:
-                            try:
-                                args = json.loads(tc.function.arguments) if tc.function.arguments else {}
-                            except json.JSONDecodeError:
-                                args = {"_raw": tc.function.arguments}
-                            tool_calls.append(LLMToolCall(id=tc.id, name=tc.function.name, arguments=args))
-
-                    content = message.content
-
-                    # Record metrics
-                    duration = time.time() - start_time
-                    usage = response.usage
-                    input_tokens = usage.prompt_tokens or 0 if usage else 0
-                    output_tokens = usage.completion_tokens or 0 if usage else 0
-
-                    metrics = get_metrics_collector()
-                    metrics.record_llm_call(
-                        provider=self.provider,
-                        model=self.model,
-                        scope=scope,
-                        duration=duration,
-                        input_tokens=input_tokens,
-                        output_tokens=output_tokens,
-                        success=True,
-                    )
-
-                    return LLMToolCallResult(
-                        content=content,
-                        tool_calls=tool_calls,
-                        finish_reason=finish_reason,
-                        input_tokens=input_tokens,
-                        output_tokens=output_tokens,
-                    )
-
-                except APIConnectionError as e:
-                    last_exception = e
-                    if attempt < max_retries:
-                        await asyncio.sleep(min(initial_backoff * (2**attempt), max_backoff))
-                        continue
-                    raise
-
-                except APIStatusError as e:
-                    if e.status_code in (401, 403):
-                        raise
-                    last_exception = e
-                    if attempt < max_retries:
-                        await asyncio.sleep(min(initial_backoff * (2**attempt), max_backoff))
-                        continue
-                    raise
-
-                except Exception:
-                    raise
-
-            if last_exception:
-                raise last_exception
-            raise RuntimeError("Tool call failed after all retries")
-
-    async def _call_with_tools_mock(
-        self,
-        messages: list[dict[str, Any]],
-        tools: list[dict[str, Any]],
-        scope: str,
-    ) -> "LLMToolCallResult":
-        """Handle mock tool calls for testing."""
-        from .response_models import LLMToolCallResult
-
-        call_record = {
-            "provider": self.provider,
-            "model": self.model,
-            "messages": messages,
-            "tools": [t.get("function", {}).get("name") for t in tools],
-            "scope": scope,
-        }
-        self._mock_calls.append(call_record)
-
-        if self._mock_response is not None:
-            if isinstance(self._mock_response, LLMToolCallResult):
-                return self._mock_response
-            # Allow setting just tool calls as a list
-            if isinstance(self._mock_response, list):
-                from .response_models import LLMToolCall
-
-                return LLMToolCallResult(
-                    tool_calls=[
-                        LLMToolCall(id=f"mock_{i}", name=tc["name"], arguments=tc.get("arguments", {}))
-                        for i, tc in enumerate(self._mock_response)
-                    ],
-                    finish_reason="tool_calls",
-                )
-
-        return LLMToolCallResult(content="mock response", finish_reason="stop")
-
-    async def _call_with_tools_anthropic(
-        self,
-        messages: list[dict[str, Any]],
-        tools: list[dict[str, Any]],
-        max_completion_tokens: int | None,
-        max_retries: int,
-        initial_backoff: float,
-        max_backoff: float,
-        start_time: float,
-        scope: str,
-    ) -> "LLMToolCallResult":
-        """Handle Anthropic tool calling."""
-        from anthropic import APIConnectionError, APIStatusError
-
-        from .response_models import LLMToolCall, LLMToolCallResult
-
-        # Convert OpenAI tool format to Anthropic format
-        anthropic_tools = []
-        for tool in tools:
-            func = tool.get("function", {})
-            anthropic_tools.append(
-                {
-                    "name": func.get("name", ""),
-                    "description": func.get("description", ""),
-                    "input_schema": func.get("parameters", {"type": "object", "properties": {}}),
-                }
-            )
-
-        # Convert messages - handle tool results
-        system_prompt = None
-        anthropic_messages = []
-        for msg in messages:
-            role = msg.get("role", "user")
-            content = msg.get("content", "")
-
-            if role == "system":
-                system_prompt = (system_prompt + "\n\n" + content) if system_prompt else content
-            elif role == "tool":
-                # Anthropic uses tool_result blocks
-                anthropic_messages.append(
-                    {
-                        "role": "user",
-                        "content": [
-                            {"type": "tool_result", "tool_use_id": msg.get("tool_call_id", ""), "content": content}
-                        ],
-                    }
-                )
-            elif role == "assistant" and msg.get("tool_calls"):
-                # Convert assistant tool calls
-                tool_use_blocks = []
-                for tc in msg["tool_calls"]:
-                    tool_use_blocks.append(
-                        {
-                            "type": "tool_use",
-                            "id": tc.get("id", ""),
-                            "name": tc.get("function", {}).get("name", ""),
-                            "input": json.loads(tc.get("function", {}).get("arguments", "{}")),
-                        }
-                    )
-                anthropic_messages.append({"role": "assistant", "content": tool_use_blocks})
-            else:
-                anthropic_messages.append({"role": role, "content": content})
-
-        call_params: dict[str, Any] = {
-            "model": self.model,
-            "messages": anthropic_messages,
-            "tools": anthropic_tools,
-            "max_tokens": max_completion_tokens or 4096,
-        }
-        if system_prompt:
-            call_params["system"] = system_prompt
-
-        last_exception = None
-        for attempt in range(max_retries + 1):
-            try:
-                response = await self._anthropic_client.messages.create(**call_params)
-
-                # Extract content and tool calls
-                content_parts = []
-                tool_calls: list[LLMToolCall] = []
-
-                for block in response.content:
-                    if block.type == "text":
-                        content_parts.append(block.text)
-                    elif block.type == "tool_use":
-                        tool_calls.append(LLMToolCall(id=block.id, name=block.name, arguments=block.input or {}))
-
-                content = "".join(content_parts) if content_parts else None
-                finish_reason = "tool_calls" if tool_calls else "stop"
-
-                # Extract token usage
-                input_tokens = response.usage.input_tokens or 0
-                output_tokens = response.usage.output_tokens or 0
-
-                # Record metrics
-                metrics = get_metrics_collector()
-                metrics.record_llm_call(
-                    provider=self.provider,
-                    model=self.model,
-                    scope=scope,
-                    duration=time.time() - start_time,
-                    input_tokens=input_tokens,
-                    output_tokens=output_tokens,
-                    success=True,
-                )
-
-                return LLMToolCallResult(
-                    content=content,
-                    tool_calls=tool_calls,
-                    finish_reason=finish_reason,
-                    input_tokens=input_tokens,
-                    output_tokens=output_tokens,
-                )
-
-            except (APIConnectionError, APIStatusError) as e:
-                if isinstance(e, APIStatusError) and e.status_code in (401, 403):
-                    raise
-                last_exception = e
-                if attempt < max_retries:
-                    await asyncio.sleep(min(initial_backoff * (2**attempt), max_backoff))
-                    continue
-                raise
-
-        if last_exception:
-            raise last_exception
-        raise RuntimeError("Anthropic tool call failed")
-
-    async def _call_with_tools_gemini(
-        self,
-        messages: list[dict[str, Any]],
-        tools: list[dict[str, Any]],
-        max_retries: int,
-        initial_backoff: float,
-        max_backoff: float,
-        start_time: float,
-        scope: str,
-    ) -> "LLMToolCallResult":
-        """Handle Gemini tool calling."""
-        from .response_models import LLMToolCall, LLMToolCallResult
-
-        # Convert tools to Gemini format
-        gemini_tools = []
-        for tool in tools:
-            func = tool.get("function", {})
-            gemini_tools.append(
-                genai_types.Tool(
-                    function_declarations=[
-                        genai_types.FunctionDeclaration(
-                            name=func.get("name", ""),
-                            description=func.get("description", ""),
-                            parameters=func.get("parameters"),
-                        )
-                    ]
-                )
-            )
-
-        # Convert messages
-        system_instruction = None
-        gemini_contents = []
-        for msg in messages:
-            role = msg.get("role", "user")
-            content = msg.get("content", "")
-
-            if role == "system":
-                system_instruction = (system_instruction + "\n\n" + content) if system_instruction else content
-            elif role == "tool":
-                # Gemini uses function_response
-                gemini_contents.append(
-                    genai_types.Content(
-                        role="user",
-                        parts=[
-                            genai_types.Part(
-                                function_response=genai_types.FunctionResponse(
-                                    name=msg.get("name", ""),
-                                    response={"result": content},
-                                )
-                            )
-                        ],
-                    )
-                )
-            elif role == "assistant":
-                gemini_contents.append(genai_types.Content(role="model", parts=[genai_types.Part(text=content)]))
-            else:
-                gemini_contents.append(genai_types.Content(role="user", parts=[genai_types.Part(text=content)]))
-
-        config = genai_types.GenerateContentConfig(
-            system_instruction=system_instruction,
-            tools=gemini_tools,
-        )
-
-        last_exception = None
-        for attempt in range(max_retries + 1):
-            try:
-                response = await self._gemini_client.aio.models.generate_content(
-                    model=self.model,
-                    contents=gemini_contents,
-                    config=config,
-                )
-
-                # Extract content and tool calls
-                content = None
-                tool_calls: list[LLMToolCall] = []
-
-                if response.candidates and response.candidates[0].content:
-                    for part in response.candidates[0].content.parts:
-                        if hasattr(part, "text") and part.text:
-                            content = part.text
-                        if hasattr(part, "function_call") and part.function_call:
-                            fc = part.function_call
-                            tool_calls.append(
-                                LLMToolCall(
-                                    id=f"gemini_{len(tool_calls)}",
-                                    name=fc.name,
-                                    arguments=dict(fc.args) if fc.args else {},
-                                )
-                            )
-
-                finish_reason = "tool_calls" if tool_calls else "stop"
-
-                # Record metrics
-                metrics = get_metrics_collector()
-                input_tokens = response.usage_metadata.prompt_token_count if response.usage_metadata else 0
-                output_tokens = response.usage_metadata.candidates_token_count if response.usage_metadata else 0
-                metrics.record_llm_call(
-                    provider=self.provider,
-                    model=self.model,
-                    scope=scope,
-                    duration=time.time() - start_time,
-                    input_tokens=input_tokens,
-                    output_tokens=output_tokens,
-                    success=True,
-                )
-
-                return LLMToolCallResult(
-                    content=content,
-                    tool_calls=tool_calls,
-                    finish_reason=finish_reason,
-                    input_tokens=input_tokens,
-                    output_tokens=output_tokens,
-                )
-
-            except genai_errors.APIError as e:
-                if e.code in (401, 403):
-                    raise
-                last_exception = e
-                if attempt < max_retries:
-                    await asyncio.sleep(min(initial_backoff * (2**attempt), max_backoff))
-                    continue
-                raise
-
-        if last_exception:
-            raise last_exception
-        raise RuntimeError("Gemini tool call failed")
-
    async def _call_anthropic(
        self,
        messages: list[dict[str, str]],
@@ -1,14 +0,0 @@
-"""
-Mental models module for Hindsight.
-
-Mental models contain directives - hard rules that are injected into reflect prompts.
-Directives are user-defined and their observations are user-provided (not LLM-generated).
-
-Other types of consolidated knowledge are handled by:
- Learnings: Automatic bottom-up consolidation from facts
- Pinned Reflections: User-curated living documents
-"""
-
-from .models import MentalModel, MentalModelSubtype
-
-__all__ = ["MentalModel", "MentalModelSubtype"]
@@ -1,53 +0,0 @@
-"""
-Pydantic models for mental models.
-"""
-
-from datetime import datetime, timezone
-from enum import Enum
-
-from pydantic import BaseModel, Field
-
-
-class MentalModelSubtype(str, Enum):
-    """Subtype of mental model.
-
-    Currently only DIRECTIVE is supported. Other types of consolidated knowledge
-    are handled by:
-    - Learnings: Automatic bottom-up consolidation from facts
-    - Pinned Reflections: User-curated living documents
-    """
-
-    DIRECTIVE = "directive"  # User-defined hard rules, observations user-provided
-
-
-class MentalModel(BaseModel):
-    """
-    A mental model representing synthesized understanding.
-
-    Mental models are the agent's consolidated knowledge. Unlike raw facts,
-    mental models provide:
-    - A one-liner description for quick scanning/retrieval
-    - A full summary for deep understanding
-    - Links to related mental models
-    """
-
-    id: str = Field(description="Unique identifier within the bank")
-    bank_id: str = Field(description="Bank this mental model belongs to")
-    subtype: MentalModelSubtype = Field(description="How this model was created")
-    name: str = Field(description="Human-readable name")
-    description: str = Field(description="One-liner for quick scanning and retrieval matching")
-    summary: str | None = Field(default=None, description="Full synthesized understanding")
-
-    # References
-    entity_id: str | None = Field(default=None, description="Reference to entities table when type=entity")
-    source_facts: list[str] = Field(default_factory=list, description="Fact IDs used to generate summary")
-    links: list[str] = Field(default_factory=list, description="Related mental model IDs")
-
-    # Tags for scoped visibility (similar to document tags)
-    tags: list[str] = Field(default_factory=list, description="Tags for scoped visibility filtering")
-
-    # Timestamps
-    last_updated: datetime | None = Field(default=None, description="When summary was last regenerated")
-    created_at: datetime = Field(
-        default_factory=lambda: datetime.now(timezone.utc), description="When this model was created"
-    )
@@ -1,18 +0,0 @@
-"""
-Reflect agent module for agentic reflection with tools.
-
-The reflect agent uses an iterative loop with tools to:
-1. Lookup mental models (existing knowledge)
-2. Recall facts (semantic + temporal search)
-3. Expand memories (get chunk/document context)
-"""
-
-from .agent import ReflectAgentResult, run_reflect_agent
-from .models import ReflectAction, ReflectActionBatch
-
-__all__ = [
-    "run_reflect_agent",
-    "ReflectAgentResult",
-    "ReflectAction",
-    "ReflectActionBatch",
-]
@@ -1,927 +0,0 @@
-"""
-Reflect agent - agentic loop for reflection with native tool calling.
-
-Uses hierarchical retrieval:
-1. search_mental_models - User-curated summaries (highest quality)
-2. search_observations - Consolidated knowledge with freshness
-3. recall - Raw facts as ground truth
-"""
-
-import asyncio
-import json
-import logging
-import re
-import time
-from typing import TYPE_CHECKING, Any, Awaitable, Callable
-
-from .models import DirectiveInfo, LLMCall, ReflectAgentResult, TokenUsageSummary, ToolCall
-from .prompts import FINAL_SYSTEM_PROMPT, _extract_directive_rules, build_final_prompt, build_system_prompt_for_tools
-from .tools_schema import get_reflect_tools
-
-
-def _build_directives_applied(directives: list[dict[str, Any]] | None) -> list[DirectiveInfo]:
-    """Build list of DirectiveInfo from directive mental models.
-
-    Handles multiple directive formats:
-    1. New format: directives have direct 'content' field
-    2. Fallback: directives have 'description' field
-    """
-    if not directives:
-        return []
-
-    result = []
-    for directive in directives:
-        directive_id = directive.get("id", "")
-        directive_name = directive.get("name", "")
-
-        # Get content from 'content' field or fallback to 'description'
-        content = directive.get("content", "") or directive.get("description", "")
-
-        result.append(DirectiveInfo(id=directive_id, name=directive_name, content=content))
-
-    return result
-
-
-if TYPE_CHECKING:
-    from ..llm_wrapper import LLMProvider
-    from ..response_models import LLMToolCall
-
-logger = logging.getLogger(__name__)
-
-DEFAULT_MAX_ITERATIONS = 10
-
-
-def _normalize_tool_name(name: str) -> str:
-    """Normalize tool name from various LLM output formats.
-
-    Some LLMs output tool names in non-standard formats:
-    - 'functions.done' (OpenAI-style prefix)
-    - 'call=functions.done' (some models)
-    - 'call=done' (some models)
-
-    Returns the normalized tool name (e.g., 'done', 'recall', etc.)
-    """
-    # Handle 'call=functions.name' or 'call=name' format
-    if name.startswith("call="):
-        name = name[len("call=") :]
-
-    # Handle 'functions.name' format
-    if name.startswith("functions."):
-        name = name[len("functions.") :]
-
-    return name
-
-
-def _is_done_tool(name: str) -> bool:
-    """Check if the tool name represents the 'done' tool."""
-    return _normalize_tool_name(name) == "done"
-
-
-# Pattern to match done() call as text - handles done({...}) with nested JSON
-_DONE_CALL_PATTERN = re.compile(r"done\s*\(\s*\{.*$", re.DOTALL)
-
-# Patterns for leaked structured output in the answer field
-_LEAKED_JSON_SUFFIX = re.compile(
-    r'\s*```(?:json)?\s*\{[^}]*(?:"(?:observation_ids|memory_ids|mental_model_ids)"|\})\s*```\s*$',
-    re.DOTALL | re.IGNORECASE,
-)
-_LEAKED_JSON_OBJECT = re.compile(
-    r'\s*\{[^{]*"(?:observation_ids|memory_ids|mental_model_ids|answer)"[^}]*\}\s*$', re.DOTALL
-)
-_TRAILING_IDS_PATTERN = re.compile(
-    r"\s*(?:observation_ids|memory_ids|mental_model_ids)\s*[=:]\s*\[.*?\]\s*$", re.DOTALL | re.IGNORECASE
-)
-
-
-def _clean_answer_text(text: str) -> str:
-    """Clean up answer text by removing any done() tool call syntax.
-
-    Some LLMs output the done() call as text instead of a proper tool call.
-    This strips out patterns like: done({"answer": "...", ...})
-    """
-    # Remove done() call pattern from the end of the text
-    cleaned = _DONE_CALL_PATTERN.sub("", text).strip()
-    return cleaned if cleaned else text
-
-
-def _clean_done_answer(text: str) -> str:
-    """Clean up the answer field from a done() tool call.
-
-    Some LLMs leak structured output patterns into the answer text, such as:
-    - JSON code blocks with observation_ids/memory_ids at the end
-    - Raw JSON objects with these fields
-    - Plain text like "observation_ids: [...]"
-
-    This cleans those patterns while preserving the actual answer content.
-    """
-    if not text:
-        return text
-
-    cleaned = text
-
-    # Remove leaked JSON in code blocks at the end
-    cleaned = _LEAKED_JSON_SUFFIX.sub("", cleaned).strip()
-
-    # Remove leaked raw JSON objects at the end
-    cleaned = _LEAKED_JSON_OBJECT.sub("", cleaned).strip()
-
-    # Remove trailing ID patterns
-    cleaned = _TRAILING_IDS_PATTERN.sub("", cleaned).strip()
-
-    return cleaned if cleaned else text
-
-
-async def _generate_structured_output(
-    answer: str,
-    response_schema: dict,
-    llm_config: "LLMProvider",
-    reflect_id: str,
-) -> tuple[dict[str, Any] | None, int, int]:
-    """Generate structured output from an answer using the provided JSON schema.
-
-    Args:
-        answer: The text answer to extract structured data from
-        response_schema: JSON Schema for the expected output structure
-        llm_config: LLM provider for making the extraction call
-        reflect_id: Reflect ID for logging
-
-    Returns:
-        Tuple of (structured_output, input_tokens, output_tokens).
-        structured_output is None if generation fails.
-    """
-    try:
-        from typing import Any as TypingAny
-
-        from pydantic import create_model
-
-        def _json_schema_type_to_python(field_schema: dict) -> type:
-            """Map JSON schema type to Python type for better LLM guidance."""
-            json_type = field_schema.get("type", "string")
-            if json_type == "array":
-                return list
-            elif json_type == "object":
-                return dict
-            elif json_type == "integer":
-                return int
-            elif json_type == "number":
-                return float
-            elif json_type == "boolean":
-                return bool
-            else:
-                return str
-
-        # Build fields from JSON schema properties
-        schema_props = response_schema.get("properties", {})
-        required_fields = set(response_schema.get("required", []))
-        fields: dict[str, TypingAny] = {}
-        for field_name, field_schema in schema_props.items():
-            field_type = _json_schema_type_to_python(field_schema)
-            default = ... if field_name in required_fields else None
-            fields[field_name] = (field_type, default)
-
-        if not fields:
-            logger.warning(f"[REFLECT {reflect_id}] No fields found in response_schema, skipping structured output")
-            return None, 0, 0
-
-        DynamicModel = create_model("StructuredResponse", **fields)
-
-        # Include the full schema in the prompt for better LLM guidance
-        schema_str = json.dumps(response_schema, indent=2)
-
-        # Build field descriptions for the prompt
-        field_descriptions = []
-        for field_name, field_schema in schema_props.items():
-            field_type = field_schema.get("type", "string")
-            field_desc = field_schema.get("description", "")
-            is_required = field_name in required_fields
-            req_marker = " (REQUIRED)" if is_required else " (optional)"
-            field_descriptions.append(f"- {field_name} ({field_type}){req_marker}: {field_desc}")
-        fields_text = "\n".join(field_descriptions)
-
-        # Call LLM with the answer to extract structured data
-        structured_prompt = f"""Your task is to extract specific information from the answer below and format it as JSON.
-
-ANSWER TO EXTRACT FROM:
-\"\"\"
-{answer}
-\"\"\"
-
-REQUIRED OUTPUT FORMAT - Extract the following fields from the answer above:
-{fields_text}
-
-JSON Schema:
-```json
-{schema_str}
-```
-
-INSTRUCTIONS:
-1. Read the answer carefully and identify the information that matches each field
-2. Extract the ACTUAL content from the answer - do NOT leave fields empty if information is present
-3. For string fields: use the exact text or a clear summary from the answer
-4. For array fields: return a JSON array (e.g., ["item1", "item2"]), NOT a string
-5. For required fields: you MUST provide a value extracted from the answer
-6. Return ONLY the JSON object, no explanation
-
-OUTPUT:"""
-
-        structured_result, usage = await llm_config.call(
-            messages=[
-                {
-                    "role": "system",
-                    "content": "You are a precise data extraction assistant. Extract information from text and return it as valid JSON matching the provided schema. Always extract actual content - never return empty strings for required fields if information is available.",
-                },
-                {"role": "user", "content": structured_prompt},
-            ],
-            response_format=DynamicModel,
-            scope="reflect_structured",
-            skip_validation=True,  # We'll handle the dict ourselves
-            return_usage=True,
-        )
-
-        # Convert to dict
-        if hasattr(structured_result, "model_dump"):
-            structured_output = structured_result.model_dump()
-        elif isinstance(structured_result, dict):
-            structured_output = structured_result
-        else:
-            # Try to parse as JSON
-            structured_output = json.loads(str(structured_result))
-
-        # Validate that required fields have non-empty values
-        for field_name in required_fields:
-            value = structured_output.get(field_name)
-            if value is None or value == "" or value == []:
-                logger.warning(f"[REFLECT {reflect_id}] Required field '{field_name}' is empty in structured output")
-
-        logger.info(f"[REFLECT {reflect_id}] Generated structured output with {len(structured_output)} fields")
-        return structured_output, usage.input_tokens, usage.output_tokens
-
-    except Exception as e:
-        logger.warning(f"[REFLECT {reflect_id}] Failed to generate structured output: {e}")
-        return None, 0, 0
-
-
-async def run_reflect_agent(
-    llm_config: "LLMProvider",
-    bank_id: str,
-    query: str,
-    bank_profile: dict[str, Any],
-    search_mental_models_fn: Callable[[str, int], Awaitable[dict[str, Any]]],
-    search_observations_fn: Callable[[str, int], Awaitable[dict[str, Any]]],
-    recall_fn: Callable[[str, int], Awaitable[dict[str, Any]]],
-    expand_fn: Callable[[list[str], str], Awaitable[dict[str, Any]]],
-    context: str | None = None,
-    max_iterations: int = DEFAULT_MAX_ITERATIONS,
-    max_tokens: int | None = None,
-    response_schema: dict | None = None,
-    directives: list[dict[str, Any]] | None = None,
-    has_mental_models: bool = False,
-    budget: str | None = None,
-) -> ReflectAgentResult:
-    """
-    Execute the reflect agent loop using native tool calling.
-
-    The agent uses hierarchical retrieval:
-    1. search_mental_models - User-curated summaries (try first)
-    2. search_observations - Consolidated knowledge with freshness
-    3. recall - Raw facts as ground truth
-
-    Args:
-        llm_config: LLM provider for agent calls
-        bank_id: Bank identifier
-        query: Question to answer
-        bank_profile: Bank profile with name and mission
-        search_mental_models_fn: Tool callback for searching mental models (query, max_results) -> result
-        search_observations_fn: Tool callback for searching observations (query, max_results) -> result
-        recall_fn: Tool callback for recall (query, max_tokens) -> result
-        expand_fn: Tool callback for expand (memory_ids, depth) -> result
-        context: Optional additional context
-        max_iterations: Maximum number of iterations before forcing response
-        max_tokens: Maximum tokens for the final response
-        response_schema: Optional JSON Schema for structured output in final response
-        directives: Optional list of directive mental models to inject as hard rules
-
-    Returns:
-        ReflectAgentResult with final answer and metadata
-    """
-    reflect_id = f"{bank_id[:8]}-{int(time.time() * 1000) % 100000}"
-    start_time = time.time()
-
-    # Build directives_applied for the trace
-    directives_applied = _build_directives_applied(directives)
-
-    # Extract directive rules for tool schema (if any)
-    directive_rules = _extract_directive_rules(directives) if directives else None
-
-    # Get tools for this agent (with directive compliance field if directives exist)
-    tools = get_reflect_tools(directive_rules=directive_rules)
-
-    # Build initial messages (directives are injected into system prompt at START and END)
-    system_prompt = build_system_prompt_for_tools(
-        bank_profile, context, directives=directives, has_mental_models=has_mental_models, budget=budget
-    )
-    messages: list[dict[str, Any]] = [
-        {"role": "system", "content": system_prompt},
-        {"role": "user", "content": query},
-    ]
-
-    # Tracking
-    total_tools_called = 0
-    tool_trace: list[ToolCall] = []
-    tool_trace_summary: list[dict[str, Any]] = []
-    llm_trace: list[dict[str, Any]] = []
-    context_history: list[dict[str, Any]] = []  # For final prompt fallback
-
-    # Token usage tracking - accumulate across all LLM calls
-    total_input_tokens = 0
-    total_output_tokens = 0
-
-    # Track available IDs for validation (prevents hallucinated citations)
-    available_memory_ids: set[str] = set()
-    available_mental_model_ids: set[str] = set()
-    available_observation_ids: set[str] = set()
-
-    def _get_llm_trace() -> list[LLMCall]:
-        return [
-            LLMCall(
-                scope=c["scope"],
-                duration_ms=c["duration_ms"],
-                input_tokens=c.get("input_tokens", 0),
-                output_tokens=c.get("output_tokens", 0),
-            )
-            for c in llm_trace
-        ]
-
-    def _get_usage() -> TokenUsageSummary:
-        return TokenUsageSummary(
-            input_tokens=total_input_tokens,
-            output_tokens=total_output_tokens,
-            total_tokens=total_input_tokens + total_output_tokens,
-        )
-
-    def _log_completion(answer: str, iterations: int, forced: bool = False):
-        elapsed_ms = int((time.time() - start_time) * 1000)
-        tools_summary = (
-            ", ".join(
-                f"{t['tool']}({t['input_summary']})={t['duration_ms']}ms/{t.get('output_chars', 0)}c"
-                for t in tool_trace_summary
-            )
-            or "none"
-        )
-        llm_summary = ", ".join(f"{c['scope']}={c['duration_ms']}ms" for c in llm_trace) or "none"
-        total_llm_ms = sum(c["duration_ms"] for c in llm_trace)
-        total_tools_ms = sum(t["duration_ms"] for t in tool_trace_summary)
-
-        answer_preview = answer[:100] + "..." if len(answer) > 100 else answer
-        mode = "forced" if forced else "done"
-        logger.info(
-            f"[REFLECT {reflect_id}] {mode} | "
-            f"query='{query[:50]}...' | "
-            f"iterations={iterations} | "
-            f"llm=[{llm_summary}] ({total_llm_ms}ms) | "
-            f"tools=[{tools_summary}] ({total_tools_ms}ms) | "
-            f"answer='{answer_preview}' | "
-            f"total={elapsed_ms}ms"
-        )
-
-    for iteration in range(max_iterations):
-        is_last = iteration == max_iterations - 1
-
-        if is_last:
-            # Force text response on last iteration - no tools
-            prompt = build_final_prompt(query, context_history, bank_profile, context)
-            llm_start = time.time()
-            response, usage = await llm_config.call(
-                messages=[
-                    {"role": "system", "content": FINAL_SYSTEM_PROMPT},
-                    {"role": "user", "content": prompt},
-                ],
-                scope="reflect_agent_final",
-                max_completion_tokens=max_tokens,
-                return_usage=True,
-            )
-            llm_duration = int((time.time() - llm_start) * 1000)
-            total_input_tokens += usage.input_tokens
-            total_output_tokens += usage.output_tokens
-            llm_trace.append(
-                {
-                    "scope": "final",
-                    "duration_ms": llm_duration,
-                    "input_tokens": usage.input_tokens,
-                    "output_tokens": usage.output_tokens,
-                }
-            )
-            answer = _clean_answer_text(response.strip())
-
-            # Generate structured output if schema provided
-            structured_output = None
-            if response_schema and answer:
-                structured_output, struct_in, struct_out = await _generate_structured_output(
-                    answer, response_schema, llm_config, reflect_id
-                )
-                total_input_tokens += struct_in
-                total_output_tokens += struct_out
-
-            _log_completion(answer, iteration + 1, forced=True)
-            return ReflectAgentResult(
-                text=answer,
-                structured_output=structured_output,
-                iterations=iteration + 1,
-                tools_called=total_tools_called,
-                tool_trace=tool_trace,
-                llm_trace=_get_llm_trace(),
-                usage=_get_usage(),
-                directives_applied=directives_applied,
-            )
-
-        # Call LLM with tools
-        llm_start = time.time()
-
-        try:
-            result = await llm_config.call_with_tools(
-                messages=messages,
-                tools=tools,
-                scope="reflect_agent",
-                tool_choice="required" if iteration == 0 else "auto",  # Force tool use on first iteration
-            )
-            llm_duration = int((time.time() - llm_start) * 1000)
-            total_input_tokens += result.input_tokens
-            total_output_tokens += result.output_tokens
-            llm_trace.append(
-                {
-                    "scope": f"agent_{iteration + 1}",
-                    "duration_ms": llm_duration,
-                    "input_tokens": result.input_tokens,
-                    "output_tokens": result.output_tokens,
-                }
-            )
-
-        except Exception as e:
-            err_duration = int((time.time() - llm_start) * 1000)
-            logger.warning(f"[REFLECT {reflect_id}] LLM error on iteration {iteration + 1}: {e} ({err_duration}ms)")
-            llm_trace.append({"scope": f"agent_{iteration + 1}_err", "duration_ms": err_duration})
-            # Guardrail: If no evidence gathered yet, retry
-            has_gathered_evidence = (
-                bool(available_memory_ids) or bool(available_mental_model_ids) or bool(available_observation_ids)
-            )
-            if not has_gathered_evidence and iteration < max_iterations - 1:
-                continue
-            prompt = build_final_prompt(query, context_history, bank_profile, context)
-            llm_start = time.time()
-            response, usage = await llm_config.call(
-                messages=[
-                    {"role": "system", "content": FINAL_SYSTEM_PROMPT},
-                    {"role": "user", "content": prompt},
-                ],
-                scope="reflect_agent_final",
-                max_completion_tokens=max_tokens,
-                return_usage=True,
-            )
-            llm_duration = int((time.time() - llm_start) * 1000)
-            total_input_tokens += usage.input_tokens
-            total_output_tokens += usage.output_tokens
-            llm_trace.append(
-                {
-                    "scope": "final",
-                    "duration_ms": llm_duration,
-                    "input_tokens": usage.input_tokens,
-                    "output_tokens": usage.output_tokens,
-                }
-            )
-            answer = _clean_answer_text(response.strip())
-
-            # Generate structured output if schema provided
-            structured_output = None
-            if response_schema and answer:
-                structured_output, struct_in, struct_out = await _generate_structured_output(
-                    answer, response_schema, llm_config, reflect_id
-                )
-                total_input_tokens += struct_in
-                total_output_tokens += struct_out
-
-            _log_completion(answer, iteration + 1, forced=True)
-            return ReflectAgentResult(
-                text=answer,
-                structured_output=structured_output,
-                iterations=iteration + 1,
-                tools_called=total_tools_called,
-                tool_trace=tool_trace,
-                llm_trace=_get_llm_trace(),
-                usage=_get_usage(),
-                directives_applied=directives_applied,
-            )
-
-        # No tool calls - LLM wants to respond with text
-        if not result.tool_calls:
-            if result.content:
-                answer = _clean_answer_text(result.content.strip())
-
-                # Generate structured output if schema provided
-                structured_output = None
-                if response_schema and answer:
-                    structured_output, struct_in, struct_out = await _generate_structured_output(
-                        answer, response_schema, llm_config, reflect_id
-                    )
-                    total_input_tokens += struct_in
-                    total_output_tokens += struct_out
-
-                _log_completion(answer, iteration + 1)
-                return ReflectAgentResult(
-                    text=answer,
-                    structured_output=structured_output,
-                    iterations=iteration + 1,
-                    tools_called=total_tools_called,
-                    tool_trace=tool_trace,
-                    llm_trace=_get_llm_trace(),
-                    usage=_get_usage(),
-                    directives_applied=directives_applied,
-                )
-            # Empty response, force final
-            prompt = build_final_prompt(query, context_history, bank_profile, context)
-            llm_start = time.time()
-            response, usage = await llm_config.call(
-                messages=[
-                    {"role": "system", "content": FINAL_SYSTEM_PROMPT},
-                    {"role": "user", "content": prompt},
-                ],
-                scope="reflect_agent_final",
-                max_completion_tokens=max_tokens,
-                return_usage=True,
-            )
-            llm_duration = int((time.time() - llm_start) * 1000)
-            total_input_tokens += usage.input_tokens
-            total_output_tokens += usage.output_tokens
-            llm_trace.append(
-                {
-                    "scope": "final",
-                    "duration_ms": llm_duration,
-                    "input_tokens": usage.input_tokens,
-                    "output_tokens": usage.output_tokens,
-                }
-            )
-            answer = _clean_answer_text(response.strip())
-
-            # Generate structured output if schema provided
-            structured_output = None
-            if response_schema and answer:
-                structured_output, struct_in, struct_out = await _generate_structured_output(
-                    answer, response_schema, llm_config, reflect_id
-                )
-                total_input_tokens += struct_in
-                total_output_tokens += struct_out
-
-            _log_completion(answer, iteration + 1, forced=True)
-            return ReflectAgentResult(
-                text=answer,
-                structured_output=structured_output,
-                iterations=iteration + 1,
-                tools_called=total_tools_called,
-                tool_trace=tool_trace,
-                llm_trace=_get_llm_trace(),
-                usage=_get_usage(),
-                directives_applied=directives_applied,
-            )
-
-        # Check for done tool call (handle various LLM output formats)
-        done_call = next((tc for tc in result.tool_calls if _is_done_tool(tc.name)), None)
-        if done_call:
-            # Guardrail: Require evidence before done
-            has_gathered_evidence = (
-                bool(available_memory_ids) or bool(available_mental_model_ids) or bool(available_observation_ids)
-            )
-            if not has_gathered_evidence and iteration < max_iterations - 1:
-                # Add assistant message and fake tool result asking for evidence
-                messages.append(
-                    {
-                        "role": "assistant",
-                        "tool_calls": [_tool_call_to_dict(done_call)],
-                    }
-                )
-                messages.append(
-                    {
-                        "role": "tool",
-                        "tool_call_id": done_call.id,
-                        "name": done_call.name,  # Required by Gemini
-                        "content": json.dumps(
-                            {
-                                "error": "You must search for information first. Use search_mental_models(), search_observations(), or recall() before providing your final answer."
-                            }
-                        ),
-                    }
-                )
-                continue
-
-            # Process done tool
-            return await _process_done_tool(
-                done_call,
-                available_memory_ids,
-                available_mental_model_ids,
-                available_observation_ids,
-                iteration + 1,
-                total_tools_called,
-                tool_trace,
-                _get_llm_trace(),
-                _get_usage(),
-                _log_completion,
-                reflect_id,
-                directives_applied=directives_applied,
-                llm_config=llm_config,
-                response_schema=response_schema,
-            )
-
-        # Execute other tools in parallel (exclude done tool in all its format variants)
-        other_tools = [tc for tc in result.tool_calls if not _is_done_tool(tc.name)]
-        if other_tools:
-            # Add assistant message with tool calls
-            messages.append(
-                {
-                    "role": "assistant",
-                    "tool_calls": [_tool_call_to_dict(tc) for tc in other_tools],
-                }
-            )
-
-            # Execute tools in parallel
-            tool_tasks = [
-                _execute_tool_with_timing(
-                    tc,
-                    search_mental_models_fn,
-                    search_observations_fn,
-                    recall_fn,
-                    expand_fn,
-                )
-                for tc in other_tools
-            ]
-            tool_results = await asyncio.gather(*tool_tasks, return_exceptions=True)
-            total_tools_called += len(other_tools)
-
-            # Process results and add to messages
-            for tc, result_data in zip(other_tools, tool_results):
-                if isinstance(result_data, Exception):
-                    # Tool execution failed - send error back to LLM so it can try again
-                    logger.warning(f"[REFLECT {reflect_id}] Tool {tc.name} failed with exception: {result_data}")
-                    output = {"error": f"Tool execution failed: {result_data}"}
-                    duration_ms = 0
-                else:
-                    output, duration_ms = result_data
-
-                # Normalize tool name for consistent tracking
-                normalized_tool_name = _normalize_tool_name(tc.name)
-
-                # Check if tool returned an error response - log but continue (LLM will see the error)
-                if isinstance(output, dict) and "error" in output:
-                    logger.warning(
-                        f"[REFLECT {reflect_id}] Tool {normalized_tool_name} returned error: {output['error']}"
-                    )
-
-                # Track available IDs from tool results (only for successful responses)
-                if (
-                    normalized_tool_name == "search_mental_models"
-                    and isinstance(output, dict)
-                    and "mental_models" in output
-                ):
-                    for mm in output["mental_models"]:
-                        if "id" in mm:
-                            available_mental_model_ids.add(mm["id"])
-
-                if (
-                    normalized_tool_name == "search_observations"
-                    and isinstance(output, dict)
-                    and "observations" in output
-                ):
-                    for obs in output["observations"]:
-                        if "id" in obs:
-                            available_observation_ids.add(obs["id"])
-
-                if normalized_tool_name == "recall" and isinstance(output, dict) and "memories" in output:
-                    for memory in output["memories"]:
-                        if "id" in memory:
-                            available_memory_ids.add(memory["id"])
-
-                # Add tool result message
-                messages.append(
-                    {
-                        "role": "tool",
-                        "tool_call_id": tc.id,
-                        "name": tc.name,  # Required by Gemini
-                        "content": json.dumps(output, default=str),
-                    }
-                )
-
-                # Track for logging and context history
-                input_dict = {"tool": tc.name, **tc.arguments}
-                input_summary = _summarize_input(tc.name, tc.arguments)
-
-                # Extract reason from tool arguments (if provided)
-                tool_reason = tc.arguments.get("reason")
-
-                tool_trace.append(
-                    ToolCall(
-                        tool=tc.name,
-                        reason=tool_reason,
-                        input=input_dict,
-                        output=output,
-                        duration_ms=duration_ms,
-                        iteration=iteration + 1,
-                    )
-                )
-
-                try:
-                    output_chars = len(json.dumps(output))
-                except (TypeError, ValueError):
-                    output_chars = len(str(output))
-
-                tool_trace_summary.append(
-                    {
-                        "tool": tc.name,
-                        "input_summary": input_summary,
-                        "duration_ms": duration_ms,
-                        "output_chars": output_chars,
-                    }
-                )
-
-                # Keep context history for fallback final prompt
-                context_history.append({"tool": tc.name, "input": input_dict, "output": output})
-
-    # Should not reach here
-    answer = "I was unable to formulate a complete answer within the iteration limit."
-    _log_completion(answer, max_iterations, forced=True)
-    return ReflectAgentResult(
-        text=answer,
-        iterations=max_iterations,
-        tools_called=total_tools_called,
-        tool_trace=tool_trace,
-        llm_trace=_get_llm_trace(),
-        usage=_get_usage(),
-        directives_applied=directives_applied,
-    )
-
-
-def _tool_call_to_dict(tc: "LLMToolCall") -> dict[str, Any]:
-    """Convert LLMToolCall to OpenAI message format."""
-    return {
-        "id": tc.id,
-        "type": "function",
-        "function": {
-            "name": tc.name,
-            "arguments": json.dumps(tc.arguments),
-        },
-    }
-
-
-async def _process_done_tool(
-    done_call: "LLMToolCall",
-    available_memory_ids: set[str],
-    available_mental_model_ids: set[str],
-    available_observation_ids: set[str],
-    iterations: int,
-    total_tools_called: int,
-    tool_trace: list[ToolCall],
-    llm_trace: list[LLMCall],
-    usage: TokenUsageSummary,
-    log_completion: Callable,
-    reflect_id: str,
-    directives_applied: list[DirectiveInfo],
-    llm_config: "LLMProvider | None" = None,
-    response_schema: dict | None = None,
-) -> ReflectAgentResult:
-    """Process the done tool call and return the result."""
-    args = done_call.arguments
-
-    # Extract and clean the answer - some LLMs leak structured output into the answer text
-    raw_answer = args.get("answer", "").strip()
-    answer = _clean_done_answer(raw_answer) if raw_answer else ""
-    if not answer:
-        answer = "No answer provided."
-
-    # Validate IDs (only include IDs that were actually retrieved)
-    used_memory_ids = [mid for mid in args.get("memory_ids", []) if mid in available_memory_ids]
-    used_mental_model_ids = [mid for mid in args.get("mental_model_ids", []) if mid in available_mental_model_ids]
-    used_observation_ids = [oid for oid in args.get("observation_ids", []) if oid in available_observation_ids]
-
-    # Generate structured output if schema provided
-    structured_output = None
-    final_usage = usage
-    if response_schema and llm_config and answer:
-        structured_output, struct_in, struct_out = await _generate_structured_output(
-            answer, response_schema, llm_config, reflect_id
-        )
-        # Add structured output tokens to usage
-        final_usage = TokenUsageSummary(
-            input_tokens=usage.input_tokens + struct_in,
-            output_tokens=usage.output_tokens + struct_out,
-            total_tokens=usage.total_tokens + struct_in + struct_out,
-        )
-
-    log_completion(answer, iterations)
-    return ReflectAgentResult(
-        text=answer,
-        structured_output=structured_output,
-        iterations=iterations,
-        tools_called=total_tools_called,
-        tool_trace=tool_trace,
-        llm_trace=llm_trace,
-        usage=final_usage,
-        used_memory_ids=used_memory_ids,
-        used_mental_model_ids=used_mental_model_ids,
-        used_observation_ids=used_observation_ids,
-        directives_applied=directives_applied,
-    )
-
-
-async def _execute_tool_with_timing(
-    tc: "LLMToolCall",
-    search_mental_models_fn: Callable[[str, int], Awaitable[dict[str, Any]]],
-    search_observations_fn: Callable[[str, int], Awaitable[dict[str, Any]]],
-    recall_fn: Callable[[str, int], Awaitable[dict[str, Any]]],
-    expand_fn: Callable[[list[str], str], Awaitable[dict[str, Any]]],
-) -> tuple[dict[str, Any], int]:
-    """Execute a tool call and return result with timing."""
-    start = time.time()
-    result = await _execute_tool(
-        tc.name,
-        tc.arguments,
-        search_mental_models_fn,
-        search_observations_fn,
-        recall_fn,
-        expand_fn,
-    )
-    duration_ms = int((time.time() - start) * 1000)
-    return result, duration_ms
-
-
-async def _execute_tool(
-    tool_name: str,
-    args: dict[str, Any],
-    search_mental_models_fn: Callable[[str, int], Awaitable[dict[str, Any]]],
-    search_observations_fn: Callable[[str, int], Awaitable[dict[str, Any]]],
-    recall_fn: Callable[[str, int], Awaitable[dict[str, Any]]],
-    expand_fn: Callable[[list[str], str], Awaitable[dict[str, Any]]],
-) -> dict[str, Any]:
-    """Execute a single tool by name."""
-    # Normalize tool name for various LLM output formats
-    tool_name = _normalize_tool_name(tool_name)
-
-    if tool_name == "search_mental_models":
-        query = args.get("query")
-        if not query:
-            return {"error": "search_mental_models requires a query parameter"}
-        max_results = args.get("max_results") or 5
-        return await search_mental_models_fn(query, max_results)
-
-    elif tool_name == "search_observations":
-        query = args.get("query")
-        if not query:
-            return {"error": "search_observations requires a query parameter"}
-        max_tokens = max(args.get("max_tokens") or 5000, 1000)  # Default 5000, min 1000
-        return await search_observations_fn(query, max_tokens)
-
-    elif tool_name == "recall":
-        query = args.get("query")
-        if not query:
-            return {"error": "recall requires a query parameter"}
-        max_tokens = max(args.get("max_tokens") or 2048, 1000)  # Default 2048, min 1000
-        return await recall_fn(query, max_tokens)
-
-    elif tool_name == "expand":
-        memory_ids = args.get("memory_ids", [])
-        if not memory_ids:
-            return {"error": "expand requires memory_ids"}
-        depth = args.get("depth", "chunk")
-        return await expand_fn(memory_ids, depth)
-
-    else:
-        return {"error": f"Unknown tool: {tool_name}"}
-
-
-def _summarize_input(tool_name: str, args: dict[str, Any]) -> str:
-    """Create a summary of tool input for logging, showing all params."""
-    if tool_name == "search_mental_models":
-        query = args.get("query", "")
-        query_preview = f"'{query[:30]}...'" if len(query) > 30 else f"'{query}'"
-        max_results = args.get("max_results") or 5
-        return f"(query={query_preview}, max_results={max_results})"
-    elif tool_name == "search_observations":
-        query = args.get("query", "")
-        query_preview = f"'{query[:30]}...'" if len(query) > 30 else f"'{query}'"
-        max_tokens = max(args.get("max_tokens") or 5000, 1000)
-        return f"(query={query_preview}, max_tokens={max_tokens})"
-    elif tool_name == "recall":
-        query = args.get("query", "")
-        query_preview = f"'{query[:30]}...'" if len(query) > 30 else f"'{query}'"
-        # Show actual value used (default 2048, min 1000)
-        max_tokens = max(args.get("max_tokens") or 2048, 1000)
-        return f"(query={query_preview}, max_tokens={max_tokens})"
-    elif tool_name == "expand":
-        memory_ids = args.get("memory_ids", [])
-        depth = args.get("depth", "chunk")
-        return f"(memory_ids=[{len(memory_ids)} ids], depth={depth})"
-    elif tool_name == "done":
-        answer = args.get("answer", "")
-        answer_preview = f"'{answer[:30]}...'" if len(answer) > 30 else f"'{answer}'"
-        memory_ids = args.get("memory_ids", [])
-        mental_model_ids = args.get("mental_model_ids", [])
-        observation_ids = args.get("observation_ids", [])
-        return (
-            f"(answer={answer_preview}, mem={len(memory_ids)}, mm={len(mental_model_ids)}, obs={len(observation_ids)})"
-        )
-    return str(args)
@@ -1,109 +0,0 @@
-"""
-Pydantic models for the reflect agent.
-"""
-
-from typing import Any, Literal
-
-from pydantic import BaseModel, Field
-
-
-class ObservationSection(BaseModel):
-    """A section within an observation with its supporting memories."""
-
-    title: str = Field(description="Section header (can be empty for intro)")
-    text: str = Field(description="Section content - no headers, use lists/tables/bold")
-    memory_ids: list[str] = Field(default_factory=list, description="Memory IDs supporting this section")
-
-
-class ReflectAction(BaseModel):
-    """Single action the reflect agent can take."""
-
-    tool: Literal["list_observations", "get_observation", "recall", "expand", "done"] = Field(
-        description="Tool to invoke: list_observations, get_observation, recall, expand, or done"
-    )
-    # Tool-specific parameters
-    observation_id: str | None = Field(default=None, description="Observation ID for get_observation")
-    query: str | None = Field(default=None, description="Search query for recall")
-    max_tokens: int | None = Field(default=None, description="Max tokens for recall results (default 2048)")
-    memory_ids: list[str] | None = Field(default=None, description="Memory unit IDs for expand (batched)")
-    depth: Literal["chunk", "document"] | None = Field(default=None, description="Expansion depth for expand")
-    observation_sections: list[ObservationSection] | None = Field(
-        default=None, description="Observation sections for done action (when output_mode=observations)"
-    )
-    # Plain text answer fields (for output_mode=answer)
-    answer: str | None = Field(default=None, description="Plain text answer for done action (no markdown)")
-    answer_memory_ids: list[str] | None = Field(
-        default=None, description="Memory IDs supporting the answer", alias="memory_ids"
-    )
-    answer_model_ids: list[str] | None = Field(
-        default=None, description="Mental model IDs supporting the answer", alias="model_ids"
-    )
-    reasoning: str | None = Field(default=None, description="Brief reasoning for this action")
-
-
-class ReflectActionBatch(BaseModel):
-    """Batch of actions for parallel execution."""
-
-    actions: list[ReflectAction] = Field(description="List of actions to execute in parallel")
-
-
-class ToolCall(BaseModel):
-    """A single tool call made during reflect."""
-
-    tool: str = Field(description="Tool name: lookup, recall, expand")
-    reason: str | None = Field(default=None, description="Agent's reasoning for making this tool call")
-    input: dict = Field(description="Tool input parameters")
-    output: dict = Field(description="Tool output/result")
-    duration_ms: int = Field(description="Execution time in milliseconds")
-    iteration: int = Field(default=0, description="Iteration number (1-based) when this tool was called")
-
-
-class LLMCall(BaseModel):
-    """A single LLM call made during reflect."""
-
-    scope: str = Field(description="Call scope: agent_1, agent_2, final, etc.")
-    duration_ms: int = Field(description="Execution time in milliseconds")
-    input_tokens: int = Field(default=0, description="Input tokens used")
-    output_tokens: int = Field(default=0, description="Output tokens used")
-
-
-class DirectiveInfo(BaseModel):
-    """Information about a directive that was applied during reflect."""
-
-    id: str = Field(description="Directive mental model ID")
-    name: str = Field(description="Directive name")
-    content: str = Field(description="Directive content")
-
-
-class TokenUsageSummary(BaseModel):
-    """Total token usage across all LLM calls."""
-
-    input_tokens: int = Field(default=0, description="Total input tokens used")
-    output_tokens: int = Field(default=0, description="Total output tokens used")
-    total_tokens: int = Field(default=0, description="Total tokens (input + output)")
-
-
-class ReflectAgentResult(BaseModel):
-    """Result from the reflect agent."""
-
-    text: str = Field(description="Final answer text")
-    structured_output: dict[str, Any] | None = Field(
-        default=None, description="Structured output parsed according to provided response_schema"
-    )
-    iterations: int = Field(default=0, description="Number of iterations taken")
-    tools_called: int = Field(default=0, description="Total number of tool calls made")
-    tool_trace: list[ToolCall] = Field(default_factory=list, description="Trace of all tool calls made")
-    llm_trace: list[LLMCall] = Field(default_factory=list, description="Trace of all LLM calls made")
-    usage: TokenUsageSummary = Field(
-        default_factory=TokenUsageSummary, description="Total token usage across all LLM calls"
-    )
-    used_memory_ids: list[str] = Field(default_factory=list, description="Validated memory IDs actually used in answer")
-    used_mental_model_ids: list[str] = Field(
-        default_factory=list, description="Validated mental model IDs actually used in answer"
-    )
-    used_observation_ids: list[str] = Field(
-        default_factory=list, description="Validated observation IDs actually used in answer"
-    )
-    directives_applied: list[DirectiveInfo] = Field(
-        default_factory=list, description="Directive mental models that affected this reflection"
-    )
@@ -1,186 +0,0 @@
-"""
-Models and utilities for evidence-grounded observations with computed trends.
-
-Observations are part of mental models and represent patterns/beliefs derived
-from memories. Each observation must be grounded in specific evidence (quotes)
-from memories, and trends are computed algorithmically from evidence timestamps.
-"""
-
-from datetime import datetime, timedelta, timezone
-from enum import Enum
-
-from pydantic import BaseModel, Field, computed_field, field_validator
-
-
-class Trend(str, Enum):
-    """Computed trend for an observation based on evidence timestamps.
-
-    Trends indicate how an observation's evidence is distributed over time:
-    - STABLE: Evidence spread across time, continues to present
-    - STRENGTHENING: More/denser evidence recently than before
-    - WEAKENING: Evidence mostly old, sparse recently
-    - NEW: All evidence within recent window
-    - STALE: No evidence in recent window (may no longer apply)
-    """
-
-    STABLE = "stable"
-    STRENGTHENING = "strengthening"
-    WEAKENING = "weakening"
-    NEW = "new"
-    STALE = "stale"
-
-
-class ObservationEvidence(BaseModel):
-    """A single piece of evidence supporting an observation.
-
-    Each evidence item must include an exact quote from the source memory
-    to ensure observations are grounded and verifiable.
-    """
-
-    memory_id: str = Field(description="ID of the memory unit this evidence comes from")
-    quote: str = Field(description="Exact quote from the memory supporting the observation")
-    relevance: str = Field(default="", description="Brief explanation of how this quote supports the observation")
-    timestamp: datetime = Field(description="When the source memory was created")
-
-    @field_validator("timestamp", mode="before")
-    @classmethod
-    def ensure_timezone_aware(cls, v: datetime | str | None) -> datetime:
-        """Ensure timestamp is always timezone-aware UTC."""
-        if v is None:
-            return datetime.now(timezone.utc)
-        if isinstance(v, str):
-            # Parse ISO format string, handling 'Z' suffix
-            v = datetime.fromisoformat(v.replace("Z", "+00:00"))
-        if isinstance(v, datetime):
-            if v.tzinfo is None:
-                return v.replace(tzinfo=timezone.utc)
-            return v
-        raise ValueError(f"Invalid timestamp type: {type(v)}")
-
-
-class Observation(BaseModel):
-    """A single observation within a mental model.
-
-    Observations represent patterns, preferences, beliefs, or other insights
-    derived from memories. Each observation must be grounded in evidence
-    with exact quotes from source memories.
-    """
-
-    title: str = Field(description="Short summary title for the observation (5-10 words)")
-    content: str = Field(description="The observation content - detailed explanation of what we believe to be true")
-    evidence: list[ObservationEvidence] = Field(default_factory=list, description="Supporting evidence with quotes")
-    created_at: datetime = Field(
-        default_factory=lambda: datetime.now(timezone.utc), description="When this observation was first created"
-    )
-
-    @field_validator("created_at", mode="before")
-    @classmethod
-    def ensure_created_at_timezone_aware(cls, v: datetime | str | None) -> datetime:
-        """Ensure created_at is always timezone-aware UTC."""
-        if v is None:
-            return datetime.now(timezone.utc)
-        if isinstance(v, str):
-            v = datetime.fromisoformat(v.replace("Z", "+00:00"))
-        if isinstance(v, datetime):
-            if v.tzinfo is None:
-                return v.replace(tzinfo=timezone.utc)
-            return v
-        raise ValueError(f"Invalid created_at type: {type(v)}")
-
-    @computed_field
-    @property
-    def trend(self) -> Trend:
-        """Compute trend from evidence timestamps."""
-        return compute_trend(self.evidence)
-
-    @computed_field
-    @property
-    def evidence_span(self) -> dict[str, str | None]:
-        """Get the time span covered by evidence."""
-        if not self.evidence:
-            return {"from": None, "to": None}
-        timestamps = [e.timestamp for e in self.evidence]
-        return {
-            "from": min(timestamps).isoformat(),
-            "to": max(timestamps).isoformat(),
-        }
-
-    @computed_field
-    @property
-    def evidence_count(self) -> int:
-        """Number of evidence items supporting this observation."""
-        return len(self.evidence)
-
-
-def compute_trend(
-    evidence: list[ObservationEvidence],
-    now: datetime | None = None,
-    recent_days: int = 30,
-    old_days: int = 90,
-) -> Trend:
-    """Compute the trend for an observation based on evidence timestamps.
-
-    The trend indicates how the evidence is distributed over time:
-    - STABLE: Evidence spread across time, continues to present
-    - STRENGTHENING: More evidence recently than historically
-    - WEAKENING: Evidence mostly old, sparse recently
-    - NEW: All evidence is recent (within recent_days)
-    - STALE: No evidence in recent window
-
-    Args:
-        evidence: List of evidence items with timestamps
-        now: Reference time for calculations (defaults to current UTC time)
-        recent_days: Number of days to consider "recent" (default 30)
-        old_days: Number of days to consider "old" (default 90)
-
-    Returns:
-        Computed Trend enum value
-    """
-    if now is None:
-        now = datetime.now(timezone.utc)
-
-    # Ensure now is timezone-aware
-    if now.tzinfo is None:
-        now = now.replace(tzinfo=timezone.utc)
-
-    if not evidence:
-        return Trend.STALE
-
-    recent_cutoff = now - timedelta(days=recent_days)
-    old_cutoff = now - timedelta(days=old_days)
-
-    # Normalize timestamps to UTC for comparison
-    def normalize_ts(ts: datetime) -> datetime:
-        if ts.tzinfo is None:
-            return ts.replace(tzinfo=timezone.utc)
-        return ts
-
-    recent = [e for e in evidence if normalize_ts(e.timestamp) > recent_cutoff]
-    old = [e for e in evidence if normalize_ts(e.timestamp) < old_cutoff]
-    middle = [e for e in evidence if old_cutoff <= normalize_ts(e.timestamp) <= recent_cutoff]
-
-    # No recent evidence = stale
-    if not recent:
-        return Trend.STALE
-
-    # All evidence is recent = new
-    if not old and not middle:
-        return Trend.NEW
-
-    # Compare density (evidence per day)
-    recent_density = len(recent) / recent_days if recent_days > 0 else 0
-    older_period = old_days - recent_days
-    older_density = (len(old) + len(middle)) / older_period if older_period > 0 else 0
-
-    # Avoid division by zero
-    if older_density == 0:
-        return Trend.NEW
-
-    ratio = recent_density / older_density
-
-    if ratio > 1.5:
-        return Trend.STRENGTHENING
-    elif ratio < 0.5:
-        return Trend.WEAKENING
-    else:
-        return Trend.STABLE
@@ -1,483 +0,0 @@
-"""
-System prompts for the reflect agent.
-
-The reflect agent uses hierarchical retrieval:
-1. search_mental_models - User-curated summaries (highest quality)
-2. search_observations - Consolidated knowledge with freshness awareness
-3. recall - Raw facts as ground truth fallback
-"""
-
-import json
-from typing import Any
-
-
-def _extract_directive_rules(directives: list[dict[str, Any]]) -> list[str]:
-    """
-    Extract directive rules as a list of strings.
-
-    Args:
-        directives: List of directives with name and content
-
-    Returns:
-        List of directive rule strings
-    """
-    rules = []
-    for directive in directives:
-        directive_name = directive.get("name", "")
-        # New format: directives have direct content field
-        content = directive.get("content", "")
-        if content:
-            if directive_name:
-                rules.append(f"**{directive_name}**: {content}")
-            else:
-                rules.append(content)
-        else:
-            # Legacy format: check for observations
-            observations = directive.get("observations", [])
-            if observations:
-                for obs in observations:
-                    # Support both Pydantic Observation objects and dicts
-                    if hasattr(obs, "title"):
-                        title = obs.title
-                        obs_content = obs.content
-                    else:
-                        title = obs.get("title", "")
-                        obs_content = obs.get("content", "")
-                    if title and obs_content:
-                        rules.append(f"**{title}**: {obs_content}")
-                    elif obs_content:
-                        rules.append(obs_content)
-            elif directive_name:
-                # Fallback to description
-                desc = directive.get("description", "")
-                if desc:
-                    rules.append(f"**{directive_name}**: {desc}")
-    return rules
-
-
-def build_directives_section(directives: list[dict[str, Any]]) -> str:
-    """
-    Build the directives section for the system prompt.
-
-    Directives are hard rules that MUST be followed in all responses.
-
-    Args:
-        directives: List of directive mental models with observations
-    """
-    if not directives:
-        return ""
-
-    rules = _extract_directive_rules(directives)
-    if not rules:
-        return ""
-
-    parts = [
-        "## DIRECTIVES (MANDATORY)",
-        "These are hard rules you MUST follow in ALL responses:",
-        "",
-    ]
-
-    for rule in rules:
-        parts.append(f"- {rule}")
-
-    parts.extend(
-        [
-            "",
-            "NEVER violate these directives, even if other context suggests otherwise.",
-            "IMPORTANT: Do NOT explain or justify how you handled directives in your answer. Just follow them silently.",
-            "",
-        ]
-    )
-    return "\n".join(parts)
-
-
-def build_directives_reminder(directives: list[dict[str, Any]]) -> str:
-    """
-    Build a reminder section for directives to place at the end of the prompt.
-
-    Args:
-        directives: List of directive mental models with observations
-    """
-    if not directives:
-        return ""
-
-    rules = _extract_directive_rules(directives)
-    if not rules:
-        return ""
-
-    parts = [
-        "",
-        "## REMINDER: MANDATORY DIRECTIVES",
-        "Before responding, ensure your answer complies with ALL of these directives:",
-        "",
-    ]
-
-    for i, rule in enumerate(rules, 1):
-        parts.append(f"{i}. {rule}")
-
-    parts.append("")
-    parts.append("Your response will be REJECTED if it violates any directive above.")
-    parts.append("Do NOT include any commentary about how you handled directives - just follow them.")
-    return "\n".join(parts)
-
-
-def build_system_prompt_for_tools(
-    bank_profile: dict[str, Any],
-    context: str | None = None,
-    directives: list[dict[str, Any]] | None = None,
-    has_mental_models: bool = False,
-    budget: str | None = None,
-) -> str:
-    """
-    Build the system prompt for tool-calling reflect agent.
-
-    The agent uses hierarchical retrieval:
-    1. search_mental_models - User-curated summaries (try first, if available)
-    2. search_observations - Consolidated knowledge with freshness
-    3. recall - Raw facts as ground truth
-
-    Args:
-        bank_profile: Bank profile with name and mission
-        context: Optional additional context
-        directives: Optional list of directive mental models to inject as hard rules
-        has_mental_models: Whether the bank has any mental models (skip if not)
-        budget: Search depth budget - "low", "mid", or "high". Controls exploration thoroughness.
-    """
-    name = bank_profile.get("name", "Assistant")
-    mission = bank_profile.get("mission", "")
-
-    parts = []
-
-    # Inject directives at the VERY START for maximum prominence
-    if directives:
-        parts.append(build_directives_section(directives))
-
-    parts.extend(
-        [
-            "You are a reflection agent that answers questions by reasoning over retrieved memories.",
-            "",
-        ]
-    )
-
-    parts.extend(
-        [
-            "## CRITICAL RULES",
-            "- You must NEVER fabricate information that has no basis in retrieved data",
-            "- You SHOULD synthesize, infer, and reason from the retrieved memories",
-            "- You MUST search before saying you don't have information",
-            "",
-            "## How to Reason",
-            "- If memories mention someone did an activity, you can infer they likely enjoyed it",
-            "- Synthesize a coherent narrative from related memories",
-            "- Be a thoughtful interpreter, not just a literal repeater",
-            "- When the exact answer isn't stated, use what IS stated to give the best answer",
-            "",
-            "## HIERARCHICAL RETRIEVAL STRATEGY",
-            "",
-        ]
-    )
-
-    # Build retrieval levels based on what's available
-    if has_mental_models:
-        parts.extend(
-            [
-                "You have access to THREE levels of knowledge. Use them in this order:",
-                "",
-                "### 1. MENTAL MODELS (search_mental_models) - Try First",
-                "- User-curated summaries about specific topics",
-                "- HIGHEST quality - manually created and maintained",
-                "- If a relevant mental model exists and is FRESH, it may fully answer the question",
-                "- Check `is_stale` field - if stale, also verify with lower levels",
-                "",
-                "### 2. OBSERVATIONS (search_observations) - Second Priority",
-                "- Auto-consolidated knowledge from memories",
-                "- Check `is_stale` field - if stale, ALSO use recall() to verify",
-                "- Good for understanding patterns and summaries",
-                "",
-                "### 3. RAW FACTS (recall) - Ground Truth",
-                "- Individual memories (world facts and experiences)",
-                "- Use when: no mental models/observations exist, they're stale, or you need specific details",
-                "- This is the source of truth that other levels are built from",
-                "",
-            ]
-        )
-    else:
-        parts.extend(
-            [
-                "You have access to TWO levels of knowledge. Use them in this order:",
-                "",
-                "### 1. OBSERVATIONS (search_observations) - Try First",
-                "- Auto-consolidated knowledge from memories",
-                "- Check `is_stale` field - if stale, ALSO use recall() to verify",
-                "- Good for understanding patterns and summaries",
-                "",
-                "### 2. RAW FACTS (recall) - Ground Truth",
-                "- Individual memories (world facts and experiences)",
-                "- Use when: no observations exist, they're stale, or you need specific details",
-                "- This is the source of truth that observations are built from",
-                "",
-            ]
-        )
-
-    parts.extend(
-        [
-            "## Query Strategy",
-            "recall() uses semantic search. NEVER just echo the user's question - decompose it into targeted searches:",
-            "",
-            "BAD: User asks 'recurring lesson themes between students' → recall('recurring lesson themes between students')",
-            "GOOD: Break it down into component searches:",
-            "  1. recall('lessons') - find all lesson-related memories",
-            "  2. recall('teaching sessions') - alternative phrasing",
-            "  3. recall('student progress') - find student-related memories",
-            "",
-            "Think: What ENTITIES and CONCEPTS does this question involve? Search for each separately.",
-            "",
-        ]
-    )
-
-    # Add budget guidance
-    if budget:
-        budget_lower = budget.lower()
-        if budget_lower == "low":
-            parts.extend(
-                [
-                    "## RESEARCH DEPTH: SHALLOW (Quick Response)",
-                    "- Prioritize speed over completeness",
-                    "- If mental models or observations provide a reasonable answer, stop there",
-                    "- Only dig deeper if the initial results are clearly insufficient",
-                    "- Prefer a quick overview rather than exhaustive details",
-                    "- Answer promptly with available information",
-                    "",
-                ]
-            )
-        elif budget_lower == "mid":
-            parts.extend(
-                [
-                    "## RESEARCH DEPTH: MODERATE (Balanced)",
-                    "- Balance thoroughness with efficiency",
-                    "- Check multiple sources when the question warrants it",
-                    "- Verify stale data if it's central to the answer",
-                    "- Don't over-explore, but ensure reasonable coverage",
-                    "",
-                ]
-            )
-        elif budget_lower == "high":
-            parts.extend(
-                [
-                    "## RESEARCH DEPTH: DEEP (Thorough Exploration)",
-                    "- Explore comprehensively before answering",
-                    "- Search across all available knowledge levels",
-                    "- Use multiple query variations to ensure coverage",
-                    "- Verify information across different retrieval levels",
-                    "- Use expand() to get full context on important memories",
-                    "- Take time to synthesize a complete, well-researched answer",
-                    "",
-                ]
-            )
-
-    parts.append("## Workflow")
-
-    if has_mental_models:
-        parts.extend(
-            [
-                "1. First, try search_mental_models() - check if a curated summary exists",
-                "2. If no mental model or it's stale, try search_observations() for consolidated knowledge",
-                "3. If observations are stale OR you need specific details, use recall() for raw facts",
-                "4. Use expand() if you need more context on specific memories",
-                "5. When ready, call done() with your answer and supporting IDs",
-            ]
-        )
-    else:
-        parts.extend(
-            [
-                "1. First, try search_observations() - check for consolidated knowledge",
-                "2. If observations are stale OR you need specific details, use recall() for raw facts",
-                "3. Use expand() if you need more context on specific memories",
-                "4. When ready, call done() with your answer and supporting IDs",
-            ]
-        )
-
-    parts.extend(
-        [
-            "",
-            "## Output Format: Plain Text Answer",
-            "Call done() with a plain text 'answer' field.",
-            "- Do NOT use markdown formatting",
-            "- NEVER include memory IDs, UUIDs, or 'Memory references' in the answer text",
-            "- Put IDs ONLY in the memory_ids/mental_model_ids/observation_ids arrays, not in the answer",
-        ]
-    )
-
-    parts.append("")
-    parts.append(f"## Memory Bank: {name}")
-
-    if mission:
-        parts.append(f"Mission: {mission}")
-
-    # Disposition traits
-    disposition = bank_profile.get("disposition", {})
-    if disposition:
-        traits = []
-        if "skepticism" in disposition:
-            traits.append(f"skepticism={disposition['skepticism']}")
-        if "literalism" in disposition:
-            traits.append(f"literalism={disposition['literalism']}")
-        if "empathy" in disposition:
-            traits.append(f"empathy={disposition['empathy']}")
-        if traits:
-            parts.append(f"Disposition: {', '.join(traits)}")
-
-    if context:
-        parts.append(f"\n## Additional Context\n{context}")
-
-    # Add directive reminder at the END for recency effect
-    if directives:
-        parts.append(build_directives_reminder(directives))
-
-    return "\n".join(parts)
-
-
-def build_agent_prompt(
-    query: str,
-    context_history: list[dict],
-    bank_profile: dict,
-    additional_context: str | None = None,
-) -> str:
-    """Build the user prompt for the reflect agent."""
-    parts = []
-
-    # Bank identity
-    name = bank_profile.get("name", "Assistant")
-    mission = bank_profile.get("mission", "")
-
-    parts.append(f"## Memory Bank Context\nName: {name}")
-    if mission:
-        parts.append(f"Mission: {mission}")
-
-    # Disposition traits if present
-    disposition = bank_profile.get("disposition", {})
-    if disposition:
-        traits = []
-        if "skepticism" in disposition:
-            traits.append(f"skepticism={disposition['skepticism']}")
-        if "literalism" in disposition:
-            traits.append(f"literalism={disposition['literalism']}")
-        if "empathy" in disposition:
-            traits.append(f"empathy={disposition['empathy']}")
-        if traits:
-            parts.append(f"Disposition: {', '.join(traits)}")
-
-    # Additional context from caller
-    if additional_context:
-        parts.append(f"\n## Additional Context\n{additional_context}")
-
-    # Tool call history
-    if context_history:
-        parts.append("\n## Tool Results (synthesize and reason from this data)")
-        for i, entry in enumerate(context_history, 1):
-            tool = entry["tool"]
-            output = entry["output"]
-            # Format as proper JSON for LLM readability
-            try:
-                output_str = json.dumps(output, indent=2, default=str)
-            except (TypeError, ValueError):
-                output_str = str(output)
-            parts.append(f"\n### Call {i}: {tool}\n```json\n{output_str}\n```")
-
-    # The question
-    parts.append(f"\n## Question\n{query}")
-
-    # Instructions
-    if context_history:
-        parts.append(
-            "\n## Instructions\n"
-            "Based on the tool results above, either call more tools or provide your final answer. "
-            "Synthesize and reason from the data - make reasonable inferences when helpful. "
-            "If you have related information, use it to give the best possible answer."
-        )
-    else:
-        parts.append(
-            "\n## Instructions\n"
-            "Start by searching for relevant information using the hierarchical retrieval strategy:\n"
-            "1. Try search_mental_models() first for curated summaries\n"
-            "2. Try search_observations() for consolidated knowledge\n"
-            "3. Use recall() for specific details or to verify stale data"
-        )
-
-    return "\n".join(parts)
-
-
-def build_final_prompt(
-    query: str,
-    context_history: list[dict],
-    bank_profile: dict,
-    additional_context: str | None = None,
-) -> str:
-    """Build the final prompt when forcing a text response (no tools)."""
-    parts = []
-
-    # Bank identity
-    name = bank_profile.get("name", "Assistant")
-    mission = bank_profile.get("mission", "")
-
-    parts.append(f"## Memory Bank Context\nName: {name}")
-    if mission:
-        parts.append(f"Mission: {mission}")
-
-    # Disposition traits if present
-    disposition = bank_profile.get("disposition", {})
-    if disposition:
-        traits = []
-        if "skepticism" in disposition:
-            traits.append(f"skepticism={disposition['skepticism']}")
-        if "literalism" in disposition:
-            traits.append(f"literalism={disposition['literalism']}")
-        if "empathy" in disposition:
-            traits.append(f"empathy={disposition['empathy']}")
-        if traits:
-            parts.append(f"Disposition: {', '.join(traits)}")
-
-    # Additional context from caller
-    if additional_context:
-        parts.append(f"\n## Additional Context\n{additional_context}")
-
-    # Tool call history
-    if context_history:
-        parts.append("\n## Retrieved Data (synthesize and reason from this data)")
-        for entry in context_history:
-            tool = entry["tool"]
-            output = entry["output"]
-            # Format as proper JSON for LLM readability
-            try:
-                output_str = json.dumps(output, indent=2, default=str)
-            except (TypeError, ValueError):
-                output_str = str(output)
-            parts.append(f"\n### From {tool}:\n```json\n{output_str}\n```")
-    else:
-        parts.append("\n## Retrieved Data\nNo data was retrieved.")
-
-    # The question
-    parts.append(f"\n## Question\n{query}")
-
-    # Final instructions
-    parts.append(
-        "\n## Instructions\n"
-        "Provide a thoughtful answer by synthesizing and reasoning from the retrieved data above. "
-        "You can make reasonable inferences from the memories, but don't completely fabricate information."
-        "If the exact answer isn't stated, use what IS stated to give the best possible answer. "
-        "Only say 'I don't have information' if the retrieved data is truly unrelated to the question."
-    )
-
-    return "\n".join(parts)
-
-
-FINAL_SYSTEM_PROMPT = """You are a thoughtful assistant that synthesizes answers from retrieved memories.
-
-Your approach:
- Reason over the retrieved memories to answer the question
- Make reasonable inferences when the exact answer isn't explicitly stated
- Connect related memories to form a complete picture
- Be helpful - if you have related information, use it to give the best possible answer
-
-Only say "I don't have information" if the retrieved data is truly unrelated to the question.
-Do NOT fabricate information that has no basis in the retrieved data."""
@@ -1,437 +0,0 @@
-"""
-Tool implementations for the reflect agent.
-
-Implements hierarchical retrieval:
-1. search_mental_models - User-curated stored reflect responses (highest quality)
-2. search_observations - Consolidated knowledge with freshness
-3. recall - Raw facts as ground truth
-"""
-
-import logging
-import uuid
-from datetime import datetime, timedelta, timezone
-from typing import TYPE_CHECKING, Any
-
-if TYPE_CHECKING:
-    from asyncpg import Connection
-
-    from ...api.http import RequestContext
-    from ..memory_engine import MemoryEngine
-
-logger = logging.getLogger(__name__)
-
-# Observation is considered stale if not updated in this many days
-STALE_THRESHOLD_DAYS = 7
-
-
-async def tool_search_mental_models(
-    conn: "Connection",
-    bank_id: str,
-    query: str,
-    query_embedding: list[float],
-    max_results: int = 5,
-    tags: list[str] | None = None,
-    tags_match: str = "any",
-    exclude_ids: list[str] | None = None,
-) -> dict[str, Any]:
-    """
-    Search user-curated mental models by semantic similarity.
-
-    Mental models are high-quality, manually created summaries about specific topics.
-    They should be searched FIRST as they represent the most reliable synthesized knowledge.
-
-    Args:
-        conn: Database connection
-        bank_id: Bank identifier
-        query: Search query (for logging/tracing)
-        query_embedding: Pre-computed embedding for semantic search
-        max_results: Maximum number of mental models to return
-        tags: Optional tags to filter mental models
-        tags_match: How to match tags - "any" (OR), "all" (AND)
-        exclude_ids: Optional list of mental model IDs to exclude (e.g., when refreshing a mental model)
-
-    Returns:
-        Dict with matching mental models including content and freshness info
-    """
-    from ..memory_engine import fq_table
-
-    # Build filters dynamically
-    filters = ""
-    params: list[Any] = [bank_id, str(query_embedding), max_results]
-    next_param = 4
-
-    if tags:
-        if tags_match == "all":
-            filters += f" AND tags @> ${next_param}::varchar[]"
-        else:
-            filters += f" AND (tags && ${next_param}::varchar[] OR tags IS NULL OR tags = '{{}}')"
-        params.append(tags)
-        next_param += 1
-
-    if exclude_ids:
-        filters += f" AND id != ALL(${next_param}::uuid[])"
-        params.append(exclude_ids)
-        next_param += 1
-
-    # Search mental models by embedding similarity
-    rows = await conn.fetch(
-        f"""
-        SELECT
-            id, name, content,
-            tags, created_at, last_refreshed_at,
-            1 - (embedding <=> $2::vector) as relevance
-        FROM {fq_table("mental_models")}
-        WHERE bank_id = $1 AND embedding IS NOT NULL {filters}
-        ORDER BY embedding <=> $2::vector
-        LIMIT $3
-        """,
-        *params,
-    )
-
-    now = datetime.now(timezone.utc)
-    mental_models = []
-
-    for row in rows:
-        last_refreshed_at = row["last_refreshed_at"]
-        if last_refreshed_at and last_refreshed_at.tzinfo is None:
-            last_refreshed_at = last_refreshed_at.replace(tzinfo=timezone.utc)
-
-        # Calculate freshness
-        is_stale = False
-        if last_refreshed_at:
-            age = now - last_refreshed_at
-            is_stale = age > timedelta(days=STALE_THRESHOLD_DAYS)
-
-        mental_models.append(
-            {
-                "id": str(row["id"]),
-                "name": row["name"],
-                "content": row["content"],
-                "tags": row["tags"] or [],
-                "relevance": round(row["relevance"], 4),
-                "updated_at": last_refreshed_at.isoformat() if last_refreshed_at else None,
-                "is_stale": is_stale,
-            }
-        )
-
-    return {
-        "query": query,
-        "count": len(mental_models),
-        "mental_models": mental_models,
-    }
-
-
-async def tool_search_observations(
-    memory_engine: "MemoryEngine",
-    bank_id: str,
-    query: str,
-    request_context: "RequestContext",
-    max_tokens: int = 5000,
-    tags: list[str] | None = None,
-    tags_match: str = "any",
-    last_consolidated_at: datetime | None = None,
-    pending_consolidation: int = 0,
-) -> dict[str, Any]:
-    """
-    Search consolidated observations using recall with include_observations.
-
-    Observations are auto-generated from memories. Returns freshness info
-    so the agent knows if it should also verify with recall().
-
-    Args:
-        memory_engine: Memory engine instance
-        bank_id: Bank identifier
-        query: Search query
-        request_context: Request context for authentication
-        max_tokens: Maximum tokens for results (default 5000)
-        tags: Optional tags to filter observations
-        tags_match: How to match tags - "any" (OR), "all" (AND)
-        last_consolidated_at: When consolidation last ran (for staleness check)
-        pending_consolidation: Number of memories waiting to be consolidated
-
-    Returns:
-        Dict with matching observations including freshness info
-    """
-    from ..memory_engine import fq_table
-
-    # Use recall to search observations (they come back in results field when fact_type=["observation"])
-    result = await memory_engine.recall_async(
-        bank_id=bank_id,
-        query=query,
-        fact_type=["observation"],  # Only retrieve observations
-        max_tokens=max_tokens,  # Token budget controls how many observations are returned
-        enable_trace=False,
-        request_context=request_context,
-        tags=tags,
-        tags_match=tags_match,
-        _connection_budget=1,
-        _quiet=True,
-    )
-
-    observations = []
-
-    # When fact_type=["observation"], results come back in `results` field as MemoryFact objects
-    # We need to fetch additional fields (proof_count, source_memory_ids) from the database
-    if result.results:
-        obs_ids = [m.id for m in result.results]
-
-        # Fetch proof_count and source_memory_ids for these observations
-        pool = await memory_engine._get_pool()
-        async with pool.acquire() as conn:
-            obs_rows = await conn.fetch(
-                f"""
-                SELECT id, proof_count, source_memory_ids
-                FROM {fq_table("memory_units")}
-                WHERE id = ANY($1::uuid[])
-                """,
-                obs_ids,
-            )
-            obs_data = {str(row["id"]): row for row in obs_rows}
-
-        for m in result.results:
-            # Get additional data from DB lookup
-            extra = obs_data.get(m.id, {})
-            proof_count = extra.get("proof_count", 1) if extra else 1
-            source_ids = extra.get("source_memory_ids", []) if extra else []
-            # Convert UUIDs to strings
-            source_memory_ids = [str(sid) for sid in (source_ids or [])]
-
-            # Determine staleness
-            is_stale = False
-            staleness_reason = None
-            if pending_consolidation > 0:
-                is_stale = True
-                staleness_reason = f"{pending_consolidation} memories pending consolidation"
-
-            observations.append(
-                {
-                    "id": str(m.id),
-                    "text": m.text,
-                    "proof_count": proof_count,
-                    "source_memory_ids": source_memory_ids,
-                    "tags": m.tags or [],
-                    "is_stale": is_stale,
-                    "staleness_reason": staleness_reason,
-                }
-            )
-
-    # Return freshness info (more understandable than raw pending_consolidation count)
-    if pending_consolidation == 0:
-        freshness = "up_to_date"
-    elif pending_consolidation < 10:
-        freshness = "slightly_stale"
-    else:
-        freshness = "stale"
-
-    return {
-        "query": query,
-        "count": len(observations),
-        "observations": observations,
-        "freshness": freshness,
-    }
-
-
-async def tool_recall(
-    memory_engine: "MemoryEngine",
-    bank_id: str,
-    query: str,
-    request_context: "RequestContext",
-    max_tokens: int = 2048,
-    max_results: int = 50,
-    tags: list[str] | None = None,
-    tags_match: str = "any",
-    connection_budget: int = 1,
-) -> dict[str, Any]:
-    """
-    Search memories using TEMPR retrieval.
-
-    This is the ground truth - raw facts and experiences.
-    Use when mental models/observations don't exist, are stale, or need verification.
-
-    Args:
-        memory_engine: Memory engine instance
-        bank_id: Bank identifier
-        query: Search query
-        request_context: Request context for authentication
-        max_tokens: Maximum tokens for results (default 2048)
-        max_results: Maximum number of results
-        tags: Filter by tags (includes untagged memories)
-        tags_match: How to match tags - "any" (OR), "all" (AND), or "exact"
-        connection_budget: Max DB connections for this recall (default 1 for internal ops)
-
-    Returns:
-        Dict with list of matching memories
-    """
-    result = await memory_engine.recall_async(
-        bank_id=bank_id,
-        query=query,
-        fact_type=["experience", "world"],  # Exclude opinions and observations
-        max_tokens=max_tokens,
-        enable_trace=False,
-        request_context=request_context,
-        tags=tags,
-        tags_match=tags_match,
-        _connection_budget=connection_budget,
-        _quiet=True,  # Suppress logging for internal operations
-    )
-
-    memories = []
-    for m in result.results[:max_results]:
-        memories.append(
-            {
-                "id": str(m.id),
-                "text": m.text,
-                "type": m.fact_type,
-                "entities": m.entities or [],
-                "occurred": m.occurred_start,  # Already ISO format string
-            }
-        )
-
-    return {
-        "query": query,
-        "count": len(memories),
-        "memories": memories,
-    }
-
-
-async def tool_expand(
-    conn: "Connection",
-    bank_id: str,
-    memory_ids: list[str],
-    depth: str,
-) -> dict[str, Any]:
-    """
-    Expand multiple memories to get chunk or document context.
-
-    Args:
-        conn: Database connection
-        bank_id: Bank identifier
-        memory_ids: List of memory unit IDs
-        depth: "chunk" or "document"
-
-    Returns:
-        Dict with results array, each containing memory, chunk, and optionally document data
-    """
-    from ..memory_engine import fq_table
-
-    if not memory_ids:
-        return {"error": "memory_ids is required and must not be empty"}
-
-    # Validate and convert UUIDs
-    valid_uuids: list[uuid.UUID] = []
-    errors: dict[str, str] = {}
-    for mid in memory_ids:
-        try:
-            valid_uuids.append(uuid.UUID(mid))
-        except ValueError:
-            errors[mid] = f"Invalid memory_id format: {mid}"
-
-    if not valid_uuids:
-        return {"error": "No valid memory IDs provided", "details": errors}
-
-    # Batch fetch all memory units
-    memories = await conn.fetch(
-        f"""
-        SELECT id, text, chunk_id, document_id, fact_type, context
-        FROM {fq_table("memory_units")}
-        WHERE id = ANY($1) AND bank_id = $2
-        """,
-        valid_uuids,
-        bank_id,
-    )
-    memory_map = {row["id"]: row for row in memories}
-
-    # Collect chunk_ids and document_ids for batch fetching
-    chunk_ids = [m["chunk_id"] for m in memories if m["chunk_id"]]
-    doc_ids_from_chunks: set[str] = set()
-    doc_ids_direct: set[str] = set()
-
-    # Batch fetch all chunks
-    chunk_map: dict[str, Any] = {}
-    if chunk_ids:
-        chunks = await conn.fetch(
-            f"""
-            SELECT chunk_id, chunk_text, chunk_index, document_id
-            FROM {fq_table("chunks")}
-            WHERE chunk_id = ANY($1)
-            """,
-            chunk_ids,
-        )
-        chunk_map = {row["chunk_id"]: row for row in chunks}
-        if depth == "document":
-            doc_ids_from_chunks = {c["document_id"] for c in chunks if c["document_id"]}
-
-    # Collect direct document IDs (memories without chunks)
-    if depth == "document":
-        for m in memories:
-            if not m["chunk_id"] and m["document_id"]:
-                doc_ids_direct.add(m["document_id"])
-
-    # Batch fetch all documents
-    doc_map: dict[str, Any] = {}
-    all_doc_ids = list(doc_ids_from_chunks | doc_ids_direct)
-    if all_doc_ids:
-        docs = await conn.fetch(
-            f"""
-            SELECT id, original_text, metadata, retain_params
-            FROM {fq_table("documents")}
-            WHERE id = ANY($1) AND bank_id = $2
-            """,
-            all_doc_ids,
-            bank_id,
-        )
-        doc_map = {row["id"]: row for row in docs}
-
-    # Build results
-    results: list[dict[str, Any]] = []
-    for mid, mem_uuid in zip(memory_ids, valid_uuids):
-        if mid in errors:
-            results.append({"memory_id": mid, "error": errors[mid]})
-            continue
-
-        memory = memory_map.get(mem_uuid)
-        if not memory:
-            results.append({"memory_id": mid, "error": f"Memory not found: {mid}"})
-            continue
-
-        item: dict[str, Any] = {
-            "memory_id": mid,
-            "memory": {
-                "id": str(memory["id"]),
-                "text": memory["text"],
-                "type": memory["fact_type"],
-                "context": memory["context"],
-            },
-        }
-
-        # Add chunk if available
-        if memory["chunk_id"] and memory["chunk_id"] in chunk_map:
-            chunk = chunk_map[memory["chunk_id"]]
-            item["chunk"] = {
-                "id": chunk["chunk_id"],
-                "text": chunk["chunk_text"],
-                "index": chunk["chunk_index"],
-                "document_id": chunk["document_id"],
-            }
-            # Add document if depth=document
-            if depth == "document" and chunk["document_id"] in doc_map:
-                doc = doc_map[chunk["document_id"]]
-                item["document"] = {
-                    "id": doc["id"],
-                    "full_text": doc["original_text"],
-                    "metadata": doc["metadata"],
-                    "retain_params": doc["retain_params"],
-                }
-        elif memory["document_id"] and depth == "document" and memory["document_id"] in doc_map:
-            # No chunk, but has document_id
-            doc = doc_map[memory["document_id"]]
-            item["document"] = {
-                "id": doc["id"],
-                "full_text": doc["original_text"],
-                "metadata": doc["metadata"],
-                "retain_params": doc["retain_params"],
-            }
-
-        results.append(item)
-
-    return {"results": results, "count": len(results)}
@@ -1,250 +0,0 @@
-"""
-Tool schema definitions for the reflect agent.
-
-These are OpenAI-format tool definitions used with native tool calling.
-The reflect agent uses a hierarchical retrieval strategy:
-1. search_mental_models - User-curated stored reflect responses (highest quality, if applicable)
-2. search_observations - Consolidated knowledge with freshness awareness
-3. recall - Raw facts (world/experience) as ground truth fallback
-"""
-
-# Tool definitions in OpenAI format
-
-TOOL_SEARCH_MENTAL_MODELS = {
-    "type": "function",
-    "function": {
-        "name": "search_mental_models",
-        "description": (
-            "Search user-curated mental models (stored reflect responses). These are high-quality, manually created "
-            "summaries about specific topics. Use FIRST when the question might be covered by an "
-            "existing mental model. Returns mental models with their content and last refresh time."
-        ),
-        "parameters": {
-            "type": "object",
-            "properties": {
-                "reason": {
-                    "type": "string",
-                    "description": "Brief explanation of why you're making this search (for debugging)",
-                },
-                "query": {
-                    "type": "string",
-                    "description": "Search query to find relevant mental models",
-                },
-                "max_results": {
-                    "type": "integer",
-                    "description": "Maximum number of mental models to return (default 5)",
-                },
-            },
-            "required": ["reason", "query"],
-        },
-    },
-}
-
-TOOL_SEARCH_OBSERVATIONS = {
-    "type": "function",
-    "function": {
-        "name": "search_observations",
-        "description": (
-            "Search consolidated observations (auto-generated knowledge). These are automatically "
-            "synthesized from memories. Returns observations with freshness info (updated_at, is_stale). "
-            "If an observation is STALE, you should ALSO use recall() to verify with current facts."
-        ),
-        "parameters": {
-            "type": "object",
-            "properties": {
-                "reason": {
-                    "type": "string",
-                    "description": "Brief explanation of why you're making this search (for debugging)",
-                },
-                "query": {
-                    "type": "string",
-                    "description": "Search query to find relevant observations",
-                },
-                "max_tokens": {
-                    "type": "integer",
-                    "description": "Maximum tokens for results (default 5000). Use higher values for broader searches.",
-                },
-            },
-            "required": ["reason", "query"],
-        },
-    },
-}
-
-TOOL_RECALL = {
-    "type": "function",
-    "function": {
-        "name": "recall",
-        "description": (
-            "Search raw memories (facts and experiences). This is the ground truth data. "
-            "Use when: (1) no reflections/mental models exist, (2) mental models are stale, "
-            "(3) you need specific details not in synthesized knowledge. "
-            "Returns individual memory facts with their timestamps."
-        ),
-        "parameters": {
-            "type": "object",
-            "properties": {
-                "reason": {
-                    "type": "string",
-                    "description": "Brief explanation of why you're making this search (for debugging)",
-                },
-                "query": {
-                    "type": "string",
-                    "description": "Search query string",
-                },
-                "max_tokens": {
-                    "type": "integer",
-                    "description": "Optional limit on result size (default 2048). Use higher values for broader searches.",
-                },
-            },
-            "required": ["reason", "query"],
-        },
-    },
-}
-
-TOOL_EXPAND = {
-    "type": "function",
-    "function": {
-        "name": "expand",
-        "description": "Get more context for one or more memories. Memory hierarchy: memory -> chunk -> document.",
-        "parameters": {
-            "type": "object",
-            "properties": {
-                "reason": {
-                    "type": "string",
-                    "description": "Brief explanation of why you need more context (for debugging)",
-                },
-                "memory_ids": {
-                    "type": "array",
-                    "items": {"type": "string"},
-                    "description": "Array of memory IDs from recall results (batch multiple for efficiency)",
-                },
-                "depth": {
-                    "type": "string",
-                    "enum": ["chunk", "document"],
-                    "description": "chunk: surrounding text chunk, document: full source document",
-                },
-            },
-            "required": ["reason", "memory_ids", "depth"],
-        },
-    },
-}
-
-TOOL_DONE_ANSWER = {
-    "type": "function",
-    "function": {
-        "name": "done",
-        "description": "Signal completion with your final answer. Use this when you have gathered enough information to answer the question.",
-        "parameters": {
-            "type": "object",
-            "properties": {
-                "answer": {
-                    "type": "string",
-                    "description": "Your response as plain text. Do NOT use markdown formatting. NEVER include memory IDs, UUIDs, or 'Memory references' in this text - put IDs only in memory_ids array.",
-                },
-                "memory_ids": {
-                    "type": "array",
-                    "items": {"type": "string"},
-                    "description": "Array of memory IDs that support your answer (put IDs here, NOT in answer text)",
-                },
-                "mental_model_ids": {
-                    "type": "array",
-                    "items": {"type": "string"},
-                    "description": "Array of mental model IDs that support your answer",
-                },
-                "observation_ids": {
-                    "type": "array",
-                    "items": {"type": "string"},
-                    "description": "Array of observation IDs that support your answer",
-                },
-            },
-            "required": ["answer"],
-        },
-    },
-}
-
-
-def _build_done_tool_with_directives(directive_rules: list[str]) -> dict:
-    """
-    Build the done tool schema with directive compliance field.
-
-    When directives are present, adds a required field that forces the agent
-    to confirm compliance with each directive before submitting.
-
-    Args:
-        directive_rules: List of directive rule strings
-    """
-    # Build rules list for description
-    rules_list = "\n".join(f"  {i + 1}. {rule}" for i, rule in enumerate(directive_rules))
-
-    # Build the tool with directive compliance field
-    return {
-        "type": "function",
-        "function": {
-            "name": "done",
-            "description": (
-                "Signal completion with your final answer. IMPORTANT: You must confirm directive compliance before submitting. "
-                "Your answer will be REJECTED if it violates any directive."
-            ),
-            "parameters": {
-                "type": "object",
-                "properties": {
-                    "answer": {
-                        "type": "string",
-                        "description": "Your response as plain text. Do NOT use markdown formatting. NEVER include memory IDs, UUIDs, or 'Memory references' in this text - put IDs only in memory_ids array.",
-                    },
-                    "memory_ids": {
-                        "type": "array",
-                        "items": {"type": "string"},
-                        "description": "Array of memory IDs that support your answer (put IDs here, NOT in answer text)",
-                    },
-                    "mental_model_ids": {
-                        "type": "array",
-                        "items": {"type": "string"},
-                        "description": "Array of mental model IDs that support your answer",
-                    },
-                    "observation_ids": {
-                        "type": "array",
-                        "items": {"type": "string"},
-                        "description": "Array of observation IDs that support your answer",
-                    },
-                    "directive_compliance": {
-                        "type": "string",
-                        "description": f"REQUIRED: Confirm your answer complies with ALL directives. List each directive and how your answer follows it:\n{rules_list}\n\nFormat: 'Directive 1: [how answer complies]. Directive 2: [how answer complies]...'",
-                    },
-                },
-                "required": ["answer", "directive_compliance"],
-            },
-        },
-    }
-
-
-def get_reflect_tools(directive_rules: list[str] | None = None) -> list[dict]:
-    """
-    Get the list of tools for the reflect agent.
-
-    The tools support a hierarchical retrieval strategy:
-    1. search_mental_models - User-curated stored reflect responses (try first)
-    2. search_observations - Consolidated knowledge with freshness
-    3. recall - Raw facts as ground truth
-
-    Args:
-        directive_rules: Optional list of directive rule strings. If provided,
-                        the done() tool will require directive compliance confirmation.
-
-    Returns:
-        List of tool definitions in OpenAI format
-    """
-    tools = [
-        TOOL_SEARCH_MENTAL_MODELS,
-        TOOL_SEARCH_OBSERVATIONS,
-        TOOL_RECALL,
-        TOOL_EXPAND,
-    ]
-
-    # Use directive-aware done tool if directives are present
-    if directive_rules:
-        tools.append(_build_done_tool_with_directives(directive_rules))
-    else:
-        tools.append(TOOL_DONE_ANSWER)
-
-    return tools
@@ -10,63 +10,8 @@ from typing import Any

 from pydantic import BaseModel, ConfigDict, Field

-# Valid fact types for recall operations (excludes 'opinion' which is deprecated)
-VALID_RECALL_FACT_TYPES = frozenset(["world", "experience", "observation"])
-
-
-class LLMToolCall(BaseModel):
-    """A tool call requested by the LLM."""
-
-    id: str = Field(description="Unique identifier for this tool call")
-    name: str = Field(description="Name of the tool to call")
-    arguments: dict[str, Any] = Field(description="Arguments to pass to the tool")
-
-
-class LLMToolCallResult(BaseModel):
-    """Result from an LLM call that may include tool calls."""
-
-    content: str | None = Field(default=None, description="Text content if any")
-    tool_calls: list[LLMToolCall] = Field(default_factory=list, description="Tool calls requested by the LLM")
-    finish_reason: str | None = Field(default=None, description="Reason the LLM stopped: 'stop', 'tool_calls', etc.")
-    input_tokens: int = Field(default=0, description="Input tokens used in this call")
-    output_tokens: int = Field(default=0, description="Output tokens used in this call")
-
-
-class ToolCallTrace(BaseModel):
-    """A single tool call made during reflect."""
-
-    tool: str = Field(description="Tool name: lookup, recall, learn, expand")
-    reason: str | None = Field(default=None, description="Agent's reasoning for making this tool call")
-    input: dict = Field(description="Tool input parameters")
-    output: dict = Field(description="Tool output/result")
-    duration_ms: int = Field(description="Execution time in milliseconds")
-    iteration: int = Field(default=0, description="Iteration number (1-based) when this tool was called")
-
-
-class LLMCallTrace(BaseModel):
-    """A single LLM call made during reflect."""
-
-    scope: str = Field(description="Call scope: agent_1, agent_2, final, etc.")
-    duration_ms: int = Field(description="Execution time in milliseconds")
-
-
-class ObservationRef(BaseModel):
-    """Reference to an observation accessed during reflect."""
-
-    id: str = Field(description="Observation ID")
-    name: str = Field(description="Observation name")
-    type: str = Field(description="Observation type: entity, concept, event")
-    subtype: str = Field(description="Observation subtype: structural, emergent, learned")
-    description: str = Field(description="Brief description")
-    summary: str | None = Field(default=None, description="Full summary (when looked up in detail)")
-
-
-class DirectiveRef(BaseModel):
-    """Reference to a directive that was applied during reflect."""
-
-    id: str = Field(description="Directive mental model ID")
-    name: str = Field(description="Directive name")
-    content: str = Field(description="Directive content")
+# Valid fact types for recall operations (excludes 'observation' which is internal)
+VALID_RECALL_FACT_TYPES = frozenset(["world", "experience", "opinion"])


 class TokenUsage(BaseModel):
@@ -140,7 +85,6 @@ class MemoryFact(BaseModel):
                "metadata": {"source": "slack"},
                "chunk_id": "bank123_session_abc123_0",
                "activation": 0.95,
-                "tags": ["user_a", "session_123"],
            }
        }
    )
@@ -158,7 +102,6 @@ class MemoryFact(BaseModel):
    chunk_id: str | None = Field(
        None, description="ID of the chunk this fact was extracted from (format: bank_id_document_id_chunk_index)"
    )
-    tags: list[str] | None = Field(None, description="Visibility scope tags associated with this fact")


 class ChunkInfo(BaseModel):
@@ -169,28 +112,6 @@ class ChunkInfo(BaseModel):
    truncated: bool = Field(default=False, description="Whether the chunk was truncated due to token limits")


-class ObservationResult(BaseModel):
-    """An observation result from recall (consolidated knowledge synthesized from facts)."""
-
-    id: str = Field(description="Unique observation ID")
-    text: str = Field(description="The observation text")
-    proof_count: int = Field(description="Number of facts supporting this observation")
-    relevance: float = Field(default=0.0, description="Relevance score to the query")
-    tags: list[str] | None = Field(default=None, description="Tags for visibility scoping")
-    source_memory_ids: list[str] = Field(
-        default_factory=list, description="IDs of facts that contribute to this observation"
-    )
-
-
-class MentalModelResult(BaseModel):
-    """A mental model result from recall (stored reflect response)."""
-
-    id: str = Field(description="Unique mental model ID")
-    name: str = Field(description="Human-readable name")
-    content: str = Field(description="The synthesized content")
-    relevance: float = Field(default=0.0, description="Relevance score to the query")
-
-
 class RecallResult(BaseModel):
    """
    Result from a recall operation.
@@ -254,14 +175,6 @@ class ReflectResult(BaseModel):
                    ],
                    "experience": [],
                    "opinion": [],
-                    "mental_models": [],
-                    "directives": [
-                        {
-                            "id": "directive-123",
-                            "name": "Response Style",
-                            "rules": ["Always be concise"],
-                        }
-                    ],
                },
                "new_opinions": ["Machine learning has great potential in healthcare"],
                "structured_output": {"summary": "ML in healthcare", "confidence": 0.9},
@@ -271,8 +184,8 @@ class ReflectResult(BaseModel):
    )

    text: str = Field(description="The formulated answer text")
-    based_on: dict[str, Any] = Field(
-        description="Facts used to formulate the answer, organized by type (world, experience, opinion, mental_models, directives)"
+    based_on: dict[str, list[MemoryFact]] = Field(
+        description="Facts used to formulate the answer, organized by type (world, experience, opinion)"
    )
    new_opinions: list[str] = Field(default_factory=list, description="List of newly formed opinions during reflection")
    structured_output: dict[str, Any] | None = Field(
@@ -283,18 +196,6 @@ class ReflectResult(BaseModel):
        default=None,
        description="Token usage metrics for the LLM calls made during this reflect operation.",
    )
-    tool_trace: list[ToolCallTrace] = Field(
-        default_factory=list,
-        description="Trace of tool calls made during reflection. Only present when include.tool_calls is enabled.",
-    )
-    llm_trace: list[LLMCallTrace] = Field(
-        default_factory=list,
-        description="Trace of LLM calls made during reflection. Only present when include.tool_calls is enabled.",
-    )
-    directives_applied: list[DirectiveRef] = Field(
-        default_factory=list,
-        description="Directive mental models that were applied during this reflection.",
-    )


 class Opinion(BaseModel):
@@ -358,32 +259,3 @@ class EntityState(BaseModel):
    observations: list[EntityObservation] = Field(
        default_factory=list, description="List of observations about this entity"
    )
-
-
-class MentalModel(BaseModel):
-    """
-    A manually configured mental model for tracking specific topics/areas.
-
-    Mental models are user-defined focus areas that the agent should track
-    and maintain summaries for, unlike auto-extracted entities.
-    """
-
-    model_config = ConfigDict(
-        json_schema_extra={
-            "example": {
-                "id": "team-dynamics",
-                "name": "Team Dynamics",
-                "description": "Track how the team collaborates, communication patterns, conflicts, and resolutions",
-                "summary": "The team has strong collaboration...",
-                "summary_updated_at": "2024-01-15T10:30:00Z",
-                "created_at": "2024-01-10T08:00:00Z",
-            }
-        }
-    )
-
-    id: str = Field(description="Unique identifier (alphanumeric lowercase)")
-    name: str = Field(description="Display name for the mental model")
-    description: str = Field(description="Prompt/directions for what to track and summarize")
-    summary: str | None = Field(None, description="Generated summary based on relevant facts")
-    summary_updated_at: str | None = Field(None, description="ISO format date when summary was last updated")
-    created_at: str = Field(description="ISO format date when the mental model was created")
@@ -1,5 +1,5 @@
 """
-bank profile utilities for disposition and mission management.
+bank profile utilities for disposition and background management.
 """

 import json
@@ -27,18 +27,19 @@ class BankProfile(TypedDict):

    name: str
    disposition: DispositionTraits
-    mission: str
+    background: str


-class MissionMergeResponse(BaseModel):
-    """LLM response for mission merge."""
+class BackgroundMergeResponse(BaseModel):
+    """LLM response for background merge with disposition inference."""

-    mission: str = Field(description="Merged mission in first person perspective")
+    background: str = Field(description="Merged background in first person perspective")
+    disposition: DispositionTraits = Field(description="Inferred disposition traits (skepticism, literalism, empathy)")


 async def get_bank_profile(pool, bank_id: str) -> BankProfile:
    """
-    Get bank profile (name, disposition + mission).
+    Get bank profile (name, disposition + background).
    Auto-creates bank with default values if not exists.

    Args:
@@ -46,13 +47,13 @@ async def get_bank_profile(pool, bank_id: str) -> BankProfile:
        bank_id: bank IDentifier

    Returns:
-        BankProfile with name, typed DispositionTraits, and mission
+        BankProfile with name, typed DispositionTraits, and background
    """
    async with acquire_with_retry(pool) as conn:
        # Try to get existing bank
        row = await conn.fetchrow(
            f"""
-            SELECT name, disposition, mission
+            SELECT name, disposition, background
            FROM {fq_table("banks")} WHERE bank_id = $1
            """,
            bank_id,
@@ -65,15 +66,13 @@ async def get_bank_profile(pool, bank_id: str) -> BankProfile:
                disposition_data = json.loads(disposition_data)

            return BankProfile(
-                name=row["name"],
-                disposition=DispositionTraits(**disposition_data),
-                mission=row["mission"] or "",
+                name=row["name"], disposition=DispositionTraits(**disposition_data), background=row["background"]
            )

        # Bank doesn't exist, create with defaults
        await conn.execute(
            f"""
-            INSERT INTO {fq_table("banks")} (bank_id, name, disposition, mission)
+            INSERT INTO {fq_table("banks")} (bank_id, name, disposition, background)
            VALUES ($1, $2, $3::jsonb, $4)
            ON CONFLICT (bank_id) DO NOTHING
            """,
@@ -83,7 +82,7 @@ async def get_bank_profile(pool, bank_id: str) -> BankProfile:
            "",
        )

-        return BankProfile(name=bank_id, disposition=DispositionTraits(**DEFAULT_DISPOSITION), mission="")
+        return BankProfile(name=bank_id, disposition=DispositionTraits(**DEFAULT_DISPOSITION), background="")


 async def update_bank_disposition(pool, bank_id: str, disposition: dict[str, int]) -> None:
@@ -111,121 +110,244 @@ async def update_bank_disposition(pool, bank_id: str, disposition: dict[str, int
        )


-async def set_bank_mission(pool, bank_id: str, mission: str) -> None:
+async def merge_bank_background(pool, llm_config, bank_id: str, new_info: str, update_disposition: bool = True) -> dict:
    """
-    Set bank mission (replacing any existing mission).
-
-    Args:
-        pool: Database connection pool
-        bank_id: bank IDentifier
-        mission: The mission text
-    """
-    # Ensure bank exists first
-    await get_bank_profile(pool, bank_id)
-
-    async with acquire_with_retry(pool) as conn:
-        await conn.execute(
-            f"""
-            UPDATE {fq_table("banks")}
-            SET mission = $2,
-                updated_at = NOW()
-            WHERE bank_id = $1
-            """,
-            bank_id,
-            mission,
-        )
-
-
-async def merge_bank_mission(pool, llm_config, bank_id: str, new_info: str) -> dict:
-    """
-    Merge new mission information with existing mission using LLM.
+    Merge new background information with existing background using LLM.
    Normalizes to first person ("I") and resolves conflicts.
+    Optionally infers disposition traits from the merged background.

    Args:
        pool: Database connection pool
-        llm_config: LLM configuration for mission merging
+        llm_config: LLM configuration for background merging
        bank_id: bank IDentifier
-        new_info: New mission information to add/merge
+        new_info: New background information to add/merge
+        update_disposition: If True, infer Big Five traits from background (default: True)

    Returns:
-        Dict with 'mission' (str) key
+        Dict with 'background' (str) and optionally 'disposition' (dict) keys
    """
    # Get current profile
    profile = await get_bank_profile(pool, bank_id)
-    current_mission = profile["mission"]
+    current_background = profile["background"]

-    # Use LLM to merge missions
-    result = await _llm_merge_mission(llm_config, current_mission, new_info)
+    # Use LLM to merge backgrounds and optionally infer disposition
+    result = await _llm_merge_background(llm_config, current_background, new_info, infer_disposition=update_disposition)

-    merged_mission = result["mission"]
+    merged_background = result["background"]
+    inferred_disposition = result.get("disposition")

    # Update in database
    async with acquire_with_retry(pool) as conn:
-        await conn.execute(
-            f"""
-            UPDATE {fq_table("banks")}
-            SET mission = $2,
-                updated_at = NOW()
-            WHERE bank_id = $1
-            """,
-            bank_id,
-            merged_mission,
-        )
+        if inferred_disposition:
+            # Update both background and disposition
+            await conn.execute(
+                f"""
+                UPDATE {fq_table("banks")}
+                SET background = $2,
+                    disposition = $3::jsonb,
+                    updated_at = NOW()
+                WHERE bank_id = $1
+                """,
+                bank_id,
+                merged_background,
+                json.dumps(inferred_disposition),
+            )
+        else:
+            # Update only background
+            await conn.execute(
+                f"""
+                UPDATE {fq_table("banks")}
+                SET background = $2,
+                    updated_at = NOW()
+                WHERE bank_id = $1
+                """,
+                bank_id,
+                merged_background,
+            )

-    return {"mission": merged_mission}
+    response = {"background": merged_background}
+    if inferred_disposition:
+        response["disposition"] = inferred_disposition
+
+    return response


-async def _llm_merge_mission(llm_config, current: str, new_info: str) -> dict:
+async def _llm_merge_background(llm_config, current: str, new_info: str, infer_disposition: bool = False) -> dict:
    """
-    Use LLM to intelligently merge mission information.
+    Use LLM to intelligently merge background information.
+    Optionally infer Big Five disposition traits from the merged background.

    Args:
        llm_config: LLM configuration to use
-        current: Current mission text
+        current: Current background text
        new_info: New information to merge
+        infer_disposition: If True, also infer disposition traits

    Returns:
-        Dict with 'mission' (str) key
+        Dict with 'background' (str) and optionally 'disposition' (dict) keys
    """
-    prompt = f"""You are helping maintain an agent's mission statement.
+    if infer_disposition:
+        prompt = f"""You are helping maintain a memory bank's background/profile and infer their disposition. You MUST respond with ONLY valid JSON.

-Current mission: {current if current else "(empty)"}
+Current background: {current if current else "(empty)"}

 New information to add: {new_info}

 Instructions:
-1. Merge the new information with the current mission
-2. If there are conflicts, the NEW information overwrites the old
+1. Merge the new information with the current background
+2. If there are conflicts (e.g., different birthplaces), the NEW information overwrites the old
+3. Keep additions that don't conflict
+4. Output in FIRST PERSON ("I") perspective
+5. Be concise - keep merged background under 500 characters
+6. Infer disposition traits from the merged background (each 1-5 integer):
+   - Skepticism: 1-5 (1=trusting, takes things at face value; 5=skeptical, questions everything)
+   - Literalism: 1-5 (1=flexible interpretation, reads between lines; 5=literal, exact interpretation)
+   - Empathy: 1-5 (1=detached, focuses on facts; 5=empathetic, considers emotional context)
+
+CRITICAL: You MUST respond with ONLY a valid JSON object. No markdown, no code blocks, no explanations. Just the JSON.
+
+Format:
+{{
+  "background": "the merged background text in first person",
+  "disposition": {{
+    "skepticism": 3,
+    "literalism": 3,
+    "empathy": 3
+  }}
+}}
+
+Trait inference examples:
+- "I'm a lawyer" → skepticism: 4, literalism: 5, empathy: 2
+- "I'm a therapist" → skepticism: 2, literalism: 2, empathy: 5
+- "I'm an engineer" → skepticism: 3, literalism: 4, empathy: 3
+- "I've been burned before by trusting people" → skepticism: 5, literalism: 3, empathy: 3
+- "I try to understand what people really mean" → skepticism: 3, literalism: 2, empathy: 4
+- "I take contracts very seriously" → skepticism: 4, literalism: 5, empathy: 2"""
+    else:
+        prompt = f"""You are helping maintain a memory bank's background/profile.
+
+Current background: {current if current else "(empty)"}
+
+New information to add: {new_info}
+
+Instructions:
+1. Merge the new information with the current background
+2. If there are conflicts (e.g., different birthplaces), the NEW information overwrites the old
 3. Keep additions that don't conflict
 4. Output in FIRST PERSON ("I") perspective
 5. Be concise - keep it under 500 characters
-6. Return ONLY the merged mission text, no explanations
+6. Return ONLY the merged background text, no explanations

-Merged mission:"""
+Merged background:"""

    try:
+        # Prepare messages
        messages = [{"role": "user", "content": prompt}]

+        if infer_disposition:
+            # Use structured output with Pydantic model for disposition inference
+            try:
+                parsed = await llm_config.call(
+                    messages=messages,
+                    response_format=BackgroundMergeResponse,
+                    scope="bank_background",
+                    temperature=0.3,
+                    max_completion_tokens=8192,
+                )
+                logger.info(f"Successfully got structured response: background={parsed.background[:100]}")
+
+                # Convert Pydantic model to dict format
+                return {"background": parsed.background, "disposition": parsed.disposition.model_dump()}
+            except Exception as e:
+                logger.warning(f"Structured output failed, falling back to manual parsing: {e}")
+                # Fall through to manual parsing below
+
+        # Manual parsing fallback or non-disposition merge
        content = await llm_config.call(
-            messages=messages, scope="bank_mission", temperature=0.3, max_completion_tokens=8192
+            messages=messages, scope="bank_background", temperature=0.3, max_completion_tokens=8192
        )

-        logger.info(f"LLM response for mission merge (first 500 chars): {content[:500]}")
+        logger.info(f"LLM response for background merge (first 500 chars): {content[:500]}")

-        merged = content.strip()
-        if not merged or merged.lower() in ["(empty)", "none", "n/a"]:
-            merged = new_info if new_info else ""
-        return {"mission": merged}
+        if infer_disposition:
+            # Parse JSON response - try multiple extraction methods
+            result = None
+
+            # Method 1: Direct parse
+            try:
+                result = json.loads(content)
+                logger.info("Successfully parsed JSON directly")
+            except json.JSONDecodeError:
+                pass
+
+            # Method 2: Extract from markdown code blocks
+            if result is None:
+                # Remove markdown code blocks
+                code_block_match = re.search(r"```(?:json)?\s*(\{.*?\})\s*```", content, re.DOTALL)
+                if code_block_match:
+                    try:
+                        result = json.loads(code_block_match.group(1))
+                        logger.info("Successfully extracted JSON from markdown code block")
+                    except json.JSONDecodeError:
+                        pass
+
+            # Method 3: Find nested JSON structure
+            if result is None:
+                # Look for JSON object with nested structure
+                json_match = re.search(
+                    r'\{[^{}]*"background"[^{}]*"disposition"[^{}]*\{[^{}]*\}[^{}]*\}', content, re.DOTALL
+                )
+                if json_match:
+                    try:
+                        result = json.loads(json_match.group())
+                        logger.info("Successfully extracted JSON using nested pattern")
+                    except json.JSONDecodeError:
+                        pass
+
+            # All parsing methods failed - use fallback
+            if result is None:
+                logger.warning(f"Failed to extract JSON from LLM response. Raw content: {content[:200]}")
+                # Fallback: use new_info as background with default disposition
+                return {
+                    "background": new_info if new_info else current if current else "",
+                    "disposition": DEFAULT_DISPOSITION.copy(),
+                }
+
+            # Validate disposition values
+            disposition = result.get("disposition", {})
+            for key in ["skepticism", "literalism", "empathy"]:
+                if key not in disposition:
+                    disposition[key] = 3  # Default to neutral
+                else:
+                    # Clamp to [1, 5] and convert to int
+                    disposition[key] = max(1, min(5, int(disposition[key])))
+
+            result["disposition"] = disposition
+
+            # Ensure background exists
+            if "background" not in result or not result["background"]:
+                result["background"] = new_info if new_info else ""
+
+            return result
+        else:
+            # Just background merge
+            merged = content
+            if not merged or merged.lower() in ["(empty)", "none", "n/a"]:
+                merged = new_info if new_info else ""
+            return {"background": merged}

    except Exception as e:
-        logger.error(f"Error merging mission with LLM: {e}")
+        logger.error(f"Error merging background with LLM: {e}")
        # Fallback: just append new info
        if current:
            merged = f"{current} {new_info}".strip()
        else:
            merged = new_info

-        return {"mission": merged}
+        result = {"background": merged}
+        if infer_disposition:
+            result["disposition"] = DEFAULT_DISPOSITION.copy()
+        return result


 async def list_banks(pool) -> list:
@@ -236,12 +358,12 @@ async def list_banks(pool) -> list:
        pool: Database connection pool

    Returns:
-        List of dicts with bank_id, name, disposition, mission, created_at, updated_at
+        List of dicts with bank_id, name, disposition, background, created_at, updated_at
    """
    async with acquire_with_retry(pool) as conn:
        rows = await conn.fetch(
            f"""
-            SELECT bank_id, name, disposition, mission, created_at, updated_at
+            SELECT bank_id, name, disposition, background, created_at, updated_at
            FROM {fq_table("banks")}
            ORDER BY updated_at DESC
            """
@@ -259,7 +381,7 @@ async def list_banks(pool) -> list:
                    "bank_id": row["bank_id"],
                    "name": row["name"],
                    "disposition": disposition_data,
-                    "mission": row["mission"] or "",
+                    "background": row["background"],
                    "created_at": row["created_at"].isoformat() if row["created_at"] else None,
                    "updated_at": row["updated_at"].isoformat() if row["updated_at"] else None,
                }
@@ -114,8 +114,11 @@ class CausalRelation(BaseModel):
    """Causal relationship from this fact to a previous fact (stored format)."""

    target_fact_index: int = Field(description="Index of the related fact in the facts array (0-based).")
-    relation_type: Literal["caused_by"] = Field(
-        description="How this fact relates to the target: 'caused_by' = this fact was caused by the target"
+    relation_type: Literal["caused_by", "enabled_by", "prevented_by"] = Field(
+        description="How this fact relates to the target: "
+        "'caused_by' = this fact was caused by the target, "
+        "'enabled_by' = this fact was enabled by the target, "
+        "'prevented_by' = this fact was prevented by the target"
    )
    strength: float = Field(
        description="Strength of relationship (0.0 to 1.0)",
@@ -138,8 +141,11 @@ class FactCausalRelation(BaseModel):
        "MUST be less than this fact's position in the list. "
        "Example: if this is fact #5, target_index can only be 0, 1, 2, 3, or 4."
    )
-    relation_type: Literal["caused_by"] = Field(
-        description="How this fact relates to the target fact: 'caused_by' = this fact was caused by the target fact"
+    relation_type: Literal["caused_by", "enabled_by", "prevented_by"] = Field(
+        description="How this fact relates to the target fact: "
+        "'caused_by' = this fact was caused by the target fact, "
+        "'enabled_by' = this fact was enabled by the target fact, "
+        "'prevented_by' = this fact was blocked/prevented by the target fact"
    )
    strength: float = Field(
        description="Strength of relationship (0.0 to 1.0). 1.0 = strong, 0.5 = moderate",
@@ -432,15 +438,34 @@ def _chunk_conversation(turns: list[dict], max_chars: int) -> list[str]:
 # FACT EXTRACTION PROMPTS
 # =============================================================================

-# Base prompt template (shared by concise and custom modes)
-# Uses {extraction_guidelines} placeholder for mode-specific instructions
-_BASE_FACT_EXTRACTION_PROMPT = """Extract SIGNIFICANT facts from text. Be SELECTIVE - only extract facts worth remembering long-term.
+# Concise extraction prompt (default) - selective, high-quality facts
+CONCISE_FACT_EXTRACTION_PROMPT = """Extract SIGNIFICANT facts from text. Be SELECTIVE - only extract facts worth remembering long-term.

-LANGUAGE REQUIREMENT: Detect the language of the input text. All extracted facts, entity names, descriptions, and other output MUST be in the SAME language as the input. Do not translate to another language.
+LANGUAGE RULE (CRITICAL): Output facts in the EXACT SAME language as the input text. If input is Japanese, output Japanese. If input is Chinese, output Chinese. NEVER translate to English. Preserve original language completely.

 {fact_types_instruction}

-{extraction_guidelines}
+══════════════════════════════════════════════════════════════════════════
+SELECTIVITY - CRITICAL (Reduces 90% of unnecessary output)
+══════════════════════════════════════════════════════════════════════════
+
+ONLY extract facts that are:
+✅ Personal info: names, relationships, roles, background
+✅ Preferences: likes, dislikes, habits, interests (e.g., "Alice likes coffee")
+✅ Significant events: milestones, decisions, achievements, changes
+✅ Plans/goals: future intentions, deadlines, commitments
+✅ Expertise: skills, knowledge, certifications, experience
+✅ Important context: projects, problems, constraints
+✅ Sensory/emotional details: feelings, sensations, perceptions that provide context
+✅ Observations: descriptions of people, places, things with specific details
+
+DO NOT extract:
+❌ Generic greetings: "how are you", "hello", pleasantries without substance
+❌ Pure filler: "thanks", "sounds good", "ok", "got it", "sure"
+❌ Process chatter: "let me check", "one moment", "I'll look into it"
+❌ Repeated info: if already stated, don't extract again
+
+CONSOLIDATE related statements into ONE fact when possible.

 ══════════════════════════════════════════════════════════════════════════
 FACT FORMAT - BE CONCISE
@@ -488,33 +513,7 @@ ENTITIES
 ══════════════════════════════════════════════════════════════════════════

 Include: people names, organizations, places, key objects, abstract concepts (career, friendship, etc.)
-Always include "user" when fact is about the user.{examples}"""
-
-# Concise mode guidelines
-_CONCISE_GUIDELINES = """══════════════════════════════════════════════════════════════════════════
-SELECTIVITY - CRITICAL (Reduces 90% of unnecessary output)
-══════════════════════════════════════════════════════════════════════════
-
-ONLY extract facts that are:
-✅ Personal info: names, relationships, roles, background
-✅ Preferences: likes, dislikes, habits, interests (e.g., "Alice likes coffee")
-✅ Significant events: milestones, decisions, achievements, changes
-✅ Plans/goals: future intentions, deadlines, commitments
-✅ Expertise: skills, knowledge, certifications, experience
-✅ Important context: projects, problems, constraints
-✅ Sensory/emotional details: feelings, sensations, perceptions that provide context
-✅ Observations: descriptions of people, places, things with specific details
-
-DO NOT extract:
-❌ Generic greetings: "how are you", "hello", pleasantries without substance
-❌ Pure filler: "thanks", "sounds good", "ok", "got it", "sure"
-❌ Process chatter: "let me check", "one moment", "I'll look into it"
-❌ Repeated info: if already stated, don't extract again
-
-CONSOLIDATE related statements into ONE fact when possible."""
-
-# Concise mode examples
-_CONCISE_EXAMPLES = """
+Always include "user" when fact is about the user.

 ══════════════════════════════════════════════════════════════════════════
 EXAMPLES
@@ -540,20 +539,6 @@ QUALITY OVER QUANTITY

 Ask: "Would this be useful to recall in 6 months?" If no, skip it."""

-# Assembled concise prompt (backward compatible - exact same output as before)
-CONCISE_FACT_EXTRACTION_PROMPT = _BASE_FACT_EXTRACTION_PROMPT.format(
-    fact_types_instruction="{fact_types_instruction}",
-    extraction_guidelines=_CONCISE_GUIDELINES,
-    examples=_CONCISE_EXAMPLES,
-)
-
-# Custom prompt uses same base but without examples
-CUSTOM_FACT_EXTRACTION_PROMPT = _BASE_FACT_EXTRACTION_PROMPT.format(
-    fact_types_instruction="{fact_types_instruction}",
-    extraction_guidelines="{custom_instructions}",
-    examples="",  # No examples for custom mode
-)
-

 # Verbose extraction prompt - detailed, comprehensive facts (legacy mode)
 VERBOSE_FACT_EXTRACTION_PROMPT = """Extract facts from text into structured format with FIVE required dimensions - BE EXTREMELY DETAILED.
@@ -677,7 +662,7 @@ CAUSAL RELATIONSHIPS
 ══════════════════════════════════════════════════════════════════════════

 Link facts with causal_relations (max 2 per fact). target_index must be < this fact's index.
-Type: "caused_by" (this fact was caused by the target fact)
+Types: "caused_by", "enabled_by", "prevented_by"

 Example: "Lost job → couldn't pay rent → moved apartment"
 - Fact 0: Lost job, causal_relations: null
@@ -701,12 +686,6 @@ async def _extract_facts_from_chunk(
    Note: event_date parameter is kept for backward compatibility but not used in prompt.
    The LLM extracts temporal information from the context string instead.
    """
-    import logging
-
-    from openai import BadRequestError
-
-    logger = logging.getLogger(__name__)
-
    memory_bank_context = f"\n- Your name: {agent_name}" if agent_name and extract_opinions else ""

    # Determine which fact types to extract based on the flag
@@ -725,27 +704,13 @@ async def _extract_facts_from_chunk(
    extract_causal_links = config.retain_extract_causal_links

    # Select base prompt based on extraction mode
-    if extraction_mode == "custom":
-        # Custom mode: inject user-provided guidelines
-        if not config.retain_custom_instructions:
-            logger.warning(
-                "extraction_mode='custom' but HINDSIGHT_API_RETAIN_CUSTOM_INSTRUCTIONS not set. "
-                "Falling back to 'concise' mode."
-            )
-            base_prompt = CONCISE_FACT_EXTRACTION_PROMPT
-            prompt = base_prompt.format(fact_types_instruction=fact_types_instruction)
-        else:
-            base_prompt = CUSTOM_FACT_EXTRACTION_PROMPT
-            prompt = base_prompt.format(
-                fact_types_instruction=fact_types_instruction,
-                custom_instructions=config.retain_custom_instructions,
-            )
-    elif extraction_mode == "verbose":
+    if extraction_mode == "verbose":
        base_prompt = VERBOSE_FACT_EXTRACTION_PROMPT
-        prompt = base_prompt.format(fact_types_instruction=fact_types_instruction)
    else:
        base_prompt = CONCISE_FACT_EXTRACTION_PROMPT
-        prompt = base_prompt.format(fact_types_instruction=fact_types_instruction)
+
+    # Format the prompt with fact types instruction
+    prompt = base_prompt.format(fact_types_instruction=fact_types_instruction)

    # Build the full prompt with or without causal relationships section
    # Select appropriate response schema based on extraction mode and causal links
@@ -758,6 +723,12 @@ async def _extract_facts_from_chunk(
    else:
        response_schema = FactExtractionResponseNoCausal

+    import logging
+
+    from openai import BadRequestError
+
+    logger = logging.getLogger(__name__)
+
    # Retry logic for JSON validation errors
    max_retries = 2
    last_error = None
@@ -852,8 +823,7 @@ Text:

                # Critical field: fact_type
                # LLM uses "assistant" but we convert to "experience" for storage
-                original_fact_type = llm_fact.get("fact_type")
-                fact_type = original_fact_type
+                fact_type = llm_fact.get("fact_type")

                # Convert "assistant" → "experience" for storage
                if fact_type == "assistant":
@@ -870,10 +840,7 @@ Text:
                    else:
                        # Default to 'world' if we can't determine
                        fact_type = "world"
-                        logger.warning(
-                            f"Fact {i}: defaulting to fact_type='world' "
-                            f"(original fact_type={original_fact_type!r}, fact_kind={fact_kind!r})"
-                        )
+                        logger.warning(f"Fact {i}: defaulting to fact_type='world'")

                # Get fact_kind for temporal handling (but don't store it)
                fact_kind = llm_fact.get("fact_kind", "conversation")
@@ -1301,7 +1268,6 @@ async def extract_facts_from_contents(
                        # mentioned_at: always the event_date (when the conversation/document occurred)
                        mentioned_at=content.event_date,
                        metadata=content.metadata,
-                        tags=content.tags,
                    )

                    extracted_facts.append(extracted_fact)
@@ -41,10 +41,10 @@ async def insert_facts_batch(
    contexts = []
    fact_types = []
    confidence_scores = []
+    access_counts = []
    metadata_jsons = []
    chunk_ids = []
    document_ids = []
-    tags_list = []

    for fact in facts:
        fact_texts.append(fact.fact_text)
@@ -60,35 +60,21 @@ async def insert_facts_batch(
        fact_types.append(fact.fact_type)
        # confidence_score is only for opinion facts
        confidence_scores.append(1.0 if fact.fact_type == "opinion" else None)
+        access_counts.append(0)  # Initial access count
        metadata_jsons.append(json.dumps(fact.metadata))
        chunk_ids.append(fact.chunk_id)
        # Use per-fact document_id if available, otherwise fallback to batch-level document_id
        document_ids.append(fact.document_id if fact.document_id else document_id)
-        # Convert tags to JSON string for proper batch insertion (PostgreSQL unnest doesn't handle 2D arrays well)
-        tags_list.append(json.dumps(fact.tags if fact.tags else []))

    # Batch insert all facts
-    # Note: tags are passed as JSON strings and converted back to varchar[] via jsonb_array_elements_text + array_agg
    results = await conn.fetch(
        f"""
-        WITH input_data AS (
-            SELECT * FROM unnest(
-                $2::text[], $3::vector[], $4::timestamptz[], $5::timestamptz[], $6::timestamptz[], $7::timestamptz[],
-                $8::text[], $9::text[], $10::float[], $11::jsonb[], $12::text[], $13::text[], $14::jsonb[]
-            ) AS t(text, embedding, event_date, occurred_start, occurred_end, mentioned_at,
-                   context, fact_type, confidence_score, metadata, chunk_id, document_id, tags_json)
-        )
        INSERT INTO {fq_table("memory_units")} (bank_id, text, embedding, event_date, occurred_start, occurred_end, mentioned_at,
-                                 context, fact_type, confidence_score, metadata, chunk_id, document_id, tags)
-        SELECT
-            $1,
-            text, embedding, event_date, occurred_start, occurred_end, mentioned_at,
-            context, fact_type, confidence_score, metadata, chunk_id, document_id,
-            COALESCE(
-                (SELECT array_agg(elem) FROM jsonb_array_elements_text(tags_json) AS elem),
-                '{{}}'::varchar[]
-            )
-        FROM input_data
+                                 context, fact_type, confidence_score, access_count, metadata, chunk_id, document_id)
+        SELECT $1, * FROM unnest(
+            $2::text[], $3::vector[], $4::timestamptz[], $5::timestamptz[], $6::timestamptz[], $7::timestamptz[],
+            $8::text[], $9::text[], $10::float[], $11::int[], $12::jsonb[], $13::text[], $14::text[]
+        )
        RETURNING id
        """,
        bank_id,
@@ -101,10 +87,10 @@ async def insert_facts_batch(
        contexts,
        fact_types,
        confidence_scores,
+        access_counts,
        metadata_jsons,
        chunk_ids,
        document_ids,
-        tags_list,
    )

    unit_ids = [str(row["id"]) for row in results]
@@ -123,7 +109,7 @@ async def ensure_bank_exists(conn, bank_id: str) -> None:
    """
    await conn.execute(
        f"""
-        INSERT INTO {fq_table("banks")} (bank_id, disposition, mission)
+        INSERT INTO {fq_table("banks")} (bank_id, disposition, background)
        VALUES ($1, $2::jsonb, $3)
        ON CONFLICT (bank_id) DO UPDATE
        SET updated_at = NOW()
@@ -135,13 +121,7 @@ async def ensure_bank_exists(conn, bank_id: str) -> None:


 async def handle_document_tracking(
-    conn,
-    bank_id: str,
-    document_id: str,
-    combined_content: str,
-    is_first_batch: bool,
-    retain_params: dict | None = None,
-    document_tags: list[str] | None = None,
+    conn, bank_id: str, document_id: str, combined_content: str, is_first_batch: bool, retain_params: dict | None = None
 ) -> None:
    """
    Handle document tracking in the database.
@@ -153,7 +133,6 @@ async def handle_document_tracking(
        combined_content: Combined content text from all content items
        is_first_batch: Whether this is the first batch (for chunked operations)
        retain_params: Optional parameters passed during retain (context, event_date, etc.)
-        document_tags: Optional list of tags to associate with the document
    """
    import hashlib

@@ -170,14 +149,13 @@ async def handle_document_tracking(
    # Insert document (or update if exists from concurrent operations)
    await conn.execute(
        f"""
-        INSERT INTO {fq_table("documents")} (id, bank_id, original_text, content_hash, metadata, retain_params, tags)
-        VALUES ($1, $2, $3, $4, $5, $6, $7)
+        INSERT INTO {fq_table("documents")} (id, bank_id, original_text, content_hash, metadata, retain_params)
+        VALUES ($1, $2, $3, $4, $5, $6)
        ON CONFLICT (id, bank_id) DO UPDATE
        SET original_text = EXCLUDED.original_text,
            content_hash = EXCLUDED.content_hash,
            metadata = EXCLUDED.metadata,
            retain_params = EXCLUDED.retain_params,
-            tags = EXCLUDED.tags,
            updated_at = NOW()
        """,
        document_id,
@@ -186,5 +164,4 @@ async def handle_document_tracking(
        content_hash,
        json.dumps({}),  # Empty metadata dict
        json.dumps(retain_params) if retain_params else None,
-        document_tags or [],
    )
@@ -754,14 +754,17 @@ async def create_causal_links_batch(
        causal_relations_per_fact: List of causal relations for each fact.
            Each element is a list of dicts with:
            - target_fact_index: Index into unit_ids for the target fact
-            - relation_type: "caused_by"
+            - relation_type: "causes", "caused_by", "enables", or "prevents"
            - strength: Float in [0.0, 1.0] representing relationship strength

    Returns:
        Number of causal links created

-    Causal link type:
-    - "caused_by": This fact was caused by the target fact
+    Causal link types:
+    - "causes": This fact directly causes the target fact (forward causation)
+    - "caused_by": This fact was caused by the target fact (backward causation)
+    - "enables": This fact enables/allows the target fact (enablement)
+    - "prevents": This fact prevents/blocks the target fact (prevention)
    """
    if not unit_ids or not causal_relations_per_fact:
        return 0
@@ -784,8 +787,8 @@ async def create_causal_links_batch(
                relation_type = relation["relation_type"]
                strength = relation.get("strength", 1.0)

-                # Validate relation_type - only "caused_by" is supported (DB constraint)
-                valid_types = {"caused_by"}
+                # Validate relation_type - must match database constraint
+                valid_types = {"causes", "caused_by", "enables", "prevents"}
                if relation_type not in valid_types:
                    logger.error(
                        f"Invalid relation_type '{relation_type}' (type: {type(relation_type).__name__}) "
@@ -0,0 +1,254 @@
+"""
+Observation regeneration for retain pipeline.
+
+Regenerates entity observations as part of the retain transaction.
+"""
+
+import logging
+import time
+import uuid
+from datetime import UTC, datetime
+
+from ...config import get_config
+from ..memory_engine import fq_table
+from ..search import observation_utils
+from . import embedding_utils
+from .types import EntityLink
+
+logger = logging.getLogger(__name__)
+
+
+def utcnow():
+    """Get current UTC time."""
+    return datetime.now(UTC)
+
+
+# Simple dataclass-like container for facts (avoid importing from memory_engine)
+class MemoryFactForObservation:
+    def __init__(self, id: str, text: str, fact_type: str, context: str, occurred_start: str | None):
+        self.id = id
+        self.text = text
+        self.fact_type = fact_type
+        self.context = context
+        self.occurred_start = occurred_start
+
+
+async def regenerate_observations_batch(
+    conn, embeddings_model, llm_config, bank_id: str, entity_links: list[EntityLink], log_buffer: list[str] = None
+) -> None:
+    """
+    Regenerate observations for top entities in this batch.
+
+    Called INSIDE the retain transaction for atomicity - if observations
+    fail, the entire retain batch is rolled back.
+
+    Args:
+        conn: Database connection (from the retain transaction)
+        embeddings_model: Embeddings model for generating observation embeddings
+        llm_config: LLM configuration for observation extraction
+        bank_id: Bank identifier
+        entity_links: Entity links from this batch
+        log_buffer: Optional log buffer for timing
+    """
+    config = get_config()
+    TOP_N_ENTITIES = config.observation_top_entities
+    MIN_FACTS_THRESHOLD = config.observation_min_facts
+
+    if not entity_links:
+        return
+
+    # Count mentions per entity in this batch
+    entity_mention_counts: dict[str, int] = {}
+    for link in entity_links:
+        if link.entity_id:
+            entity_id = str(link.entity_id)
+            entity_mention_counts[entity_id] = entity_mention_counts.get(entity_id, 0) + 1
+
+    if not entity_mention_counts:
+        return
+
+    # Sort by mention count descending and take top N
+    sorted_entities = sorted(entity_mention_counts.items(), key=lambda x: x[1], reverse=True)
+    entities_to_process = [e[0] for e in sorted_entities[:TOP_N_ENTITIES]]
+
+    obs_start = time.time()
+
+    # Convert to UUIDs
+    entity_uuids = [uuid.UUID(eid) if isinstance(eid, str) else eid for eid in entities_to_process]
+
+    # Batch query for entity names
+    entity_rows = await conn.fetch(
+        f"""
+        SELECT id, canonical_name FROM {fq_table("entities")}
+        WHERE id = ANY($1) AND bank_id = $2
+        """,
+        entity_uuids,
+        bank_id,
+    )
+    entity_names = {row["id"]: row["canonical_name"] for row in entity_rows}
+
+    # Batch query for fact counts
+    fact_counts = await conn.fetch(
+        f"""
+        SELECT ue.entity_id, COUNT(*) as cnt
+        FROM {fq_table("unit_entities")} ue
+        JOIN {fq_table("memory_units")} mu ON ue.unit_id = mu.id
+        WHERE ue.entity_id = ANY($1) AND mu.bank_id = $2
+        GROUP BY ue.entity_id
+        """,
+        entity_uuids,
+        bank_id,
+    )
+    entity_fact_counts = {row["entity_id"]: row["cnt"] for row in fact_counts}
+
+    # Filter entities that meet the threshold
+    entities_with_names = []
+    for entity_id in entities_to_process:
+        entity_uuid = uuid.UUID(entity_id) if isinstance(entity_id, str) else entity_id
+        if entity_uuid not in entity_names:
+            continue
+        fact_count = entity_fact_counts.get(entity_uuid, 0)
+        if fact_count >= MIN_FACTS_THRESHOLD:
+            entities_with_names.append((entity_id, entity_names[entity_uuid]))
+
+    if not entities_with_names:
+        return
+
+    # Process entities SEQUENTIALLY (asyncpg doesn't allow concurrent queries on same connection)
+    # We must use the same connection to stay in the retain transaction
+    total_observations = 0
+
+    for entity_id, entity_name in entities_with_names:
+        try:
+            obs_ids = await _regenerate_entity_observations(
+                conn, embeddings_model, llm_config, bank_id, entity_id, entity_name
+            )
+            total_observations += len(obs_ids)
+        except Exception as e:
+            logger.error(f"[OBSERVATIONS] Error processing entity {entity_id}: {e}")
+
+    obs_time = time.time() - obs_start
+    if log_buffer is not None:
+        log_buffer.append(
+            f"[11] Observations: {total_observations} observations for {len(entities_with_names)} entities in {obs_time:.3f}s"
+        )
+
+
+async def _regenerate_entity_observations(
+    conn, embeddings_model, llm_config, bank_id: str, entity_id: str, entity_name: str
+) -> list[str]:
+    """
+    Regenerate observations for a single entity.
+
+    Uses the provided connection (part of retain transaction).
+
+    Args:
+        conn: Database connection (from the retain transaction)
+        embeddings_model: Embeddings model
+        llm_config: LLM configuration
+        bank_id: Bank identifier
+        entity_id: Entity UUID
+        entity_name: Canonical name of the entity
+
+    Returns:
+        List of created observation IDs
+    """
+    entity_uuid = uuid.UUID(entity_id) if isinstance(entity_id, str) else entity_id
+
+    # Get all facts mentioning this entity (exclude observations themselves)
+    rows = await conn.fetch(
+        f"""
+        SELECT mu.id, mu.text, mu.context, mu.occurred_start, mu.fact_type
+        FROM {fq_table("memory_units")} mu
+        JOIN {fq_table("unit_entities")} ue ON mu.id = ue.unit_id
+        WHERE mu.bank_id = $1
+          AND ue.entity_id = $2
+          AND mu.fact_type IN ('world', 'experience')
+        ORDER BY mu.occurred_start DESC
+        LIMIT 50
+        """,
+        bank_id,
+        entity_uuid,
+    )
+
+    if not rows:
+        return []
+
+    # Convert to fact objects for observation extraction
+    facts = []
+    for row in rows:
+        occurred_start = row["occurred_start"].isoformat() if row["occurred_start"] else None
+        facts.append(
+            MemoryFactForObservation(
+                id=str(row["id"]),
+                text=row["text"],
+                fact_type=row["fact_type"],
+                context=row["context"],
+                occurred_start=occurred_start,
+            )
+        )
+
+    # Extract observations using LLM
+    observations = await observation_utils.extract_observations_from_facts(llm_config, entity_name, facts)
+
+    if not observations:
+        return []
+
+    # Delete old observations for this entity
+    await conn.execute(
+        f"""
+        DELETE FROM {fq_table("memory_units")}
+        WHERE id IN (
+            SELECT mu.id
+            FROM {fq_table("memory_units")} mu
+            JOIN {fq_table("unit_entities")} ue ON mu.id = ue.unit_id
+            WHERE mu.bank_id = $1
+              AND mu.fact_type = 'observation'
+              AND ue.entity_id = $2
+        )
+        """,
+        bank_id,
+        entity_uuid,
+    )
+
+    # Generate embeddings for new observations
+    embeddings = await embedding_utils.generate_embeddings_batch(embeddings_model, observations)
+
+    # Insert new observations
+    current_time = utcnow()
+    created_ids = []
+
+    for obs_text, embedding in zip(observations, embeddings):
+        result = await conn.fetchrow(
+            f"""
+            INSERT INTO {fq_table("memory_units")} (
+                bank_id, text, embedding, context, event_date,
+                occurred_start, occurred_end, mentioned_at,
+                fact_type, access_count
+            )
+            VALUES ($1, $2, $3, $4, $5, $6, $7, $8, 'observation', 0)
+            RETURNING id
+            """,
+            bank_id,
+            obs_text,
+            str(embedding),
+            f"observation about {entity_name}",
+            current_time,
+            current_time,
+            current_time,
+            current_time,
+        )
+        obs_id = str(result["id"])
+        created_ids.append(obs_id)
+
+        # Link observation to entity
+        await conn.execute(
+            f"""
+            INSERT INTO {fq_table("unit_entities")} (unit_id, entity_id)
+            VALUES ($1, $2)
+            """,
+            uuid.UUID(obs_id),
+            entity_uuid,
+        )
+
+    return created_ids
@@ -9,6 +9,7 @@ import time
 import uuid
 from datetime import UTC, datetime

+from ...config import get_config
 from ..db_utils import acquire_with_retry
 from . import bank_utils

@@ -27,8 +28,9 @@ from . import (
    fact_extraction,
    fact_storage,
    link_creation,
+    observation_regeneration,
 )
-from .types import EntityLink, ExtractedFact, ProcessedFact, RetainContent, RetainContentDict
+from .types import ExtractedFact, ProcessedFact, RetainContent, RetainContentDict

 logger = logging.getLogger(__name__)

@@ -38,6 +40,7 @@ async def retain_batch(
    embeddings_model,
    llm_config,
    entity_resolver,
+    task_backend,
    format_date_fn,
    duplicate_checker_fn,
    bank_id: str,
@@ -46,7 +49,6 @@ async def retain_batch(
    is_first_batch: bool = True,
    fact_type_override: str | None = None,
    confidence_score: float | None = None,
-    document_tags: list[str] | None = None,
 ) -> tuple[list[list[str]], TokenUsage]:
    """
    Process a batch of content through the retain pipeline.
@@ -56,6 +58,7 @@ async def retain_batch(
        embeddings_model: Embeddings model for generating embeddings
        llm_config: LLM configuration for fact extraction
        entity_resolver: Entity resolver for entity processing
+        task_backend: Task backend for background jobs
        format_date_fn: Function to format datetime to readable string
        duplicate_checker_fn: Function to check for duplicate facts
        bank_id: Bank identifier
@@ -64,7 +67,6 @@ async def retain_batch(
        is_first_batch: Whether this is the first batch
        fact_type_override: Override fact type for all facts
        confidence_score: Confidence score for opinions
-        document_tags: Tags applied to all items in this batch

    Returns:
        Tuple of (unit ID lists, token usage for fact extraction)
@@ -86,16 +88,12 @@ async def retain_batch(
    # Convert dicts to RetainContent objects
    contents = []
    for item in contents_dicts:
-        # Merge item-level tags with document-level tags
-        item_tags = item.get("tags", []) or []
-        merged_tags = list(set(item_tags + (document_tags or [])))
        content = RetainContent(
            content=item["content"],
            context=item.get("context", ""),
            event_date=item.get("event_date") or utcnow(),
            metadata=item.get("metadata", {}),
            entities=item.get("entities", []),
-            tags=merged_tags,
        )
        contents.append(content)

@@ -133,7 +131,7 @@ async def retain_batch(
                        if first_item.get("metadata"):
                            retain_params["metadata"] = first_item["metadata"]
                    await fact_storage.handle_document_tracking(
-                        conn, bank_id, document_id, combined_content, is_first_batch, retain_params, document_tags
+                        conn, bank_id, document_id, combined_content, is_first_batch, retain_params
                    )
                else:
                    # Check for per-item document_ids
@@ -161,7 +159,7 @@ async def retain_batch(
                            if first_item.get("metadata"):
                                retain_params["metadata"] = first_item["metadata"]
                        await fact_storage.handle_document_tracking(
-                            conn, bank_id, doc_id, combined_content, is_first_batch, retain_params, document_tags
+                            conn, bank_id, doc_id, combined_content, is_first_batch, retain_params
                        )

        total_time = time.time() - start_time
@@ -227,7 +225,7 @@ async def retain_batch(
                        retain_params["metadata"] = first_item["metadata"]

                await fact_storage.handle_document_tracking(
-                    conn, bank_id, document_id, combined_content, is_first_batch, retain_params, document_tags
+                    conn, bank_id, document_id, combined_content, is_first_batch, retain_params
                )
                document_ids_added.append(document_id)
                doc_id_mapping[None] = document_id  # For backwards compatibility
@@ -271,13 +269,7 @@ async def retain_batch(
                                    retain_params["metadata"] = first_item["metadata"]

                            await fact_storage.handle_document_tracking(
-                                conn,
-                                bank_id,
-                                actual_doc_id,
-                                combined_content,
-                                is_first_batch,
-                                retain_params,
-                                document_tags,
+                                conn, bank_id, actual_doc_id, combined_content, is_first_batch, retain_params
                            )
                            document_ids_added.append(actual_doc_id)

@@ -404,9 +396,27 @@ async def retain_batch(
            causal_link_count = await link_creation.create_causal_links_batch(conn, unit_ids, non_duplicate_facts)
            log_buffer.append(f"[10] Causal links: {causal_link_count} links in {time.time() - step_start:.3f}s")

+            # Regenerate observations - sync (in transaction) or async (background task)
+            config = get_config()
+            if config.retain_observations_async:
+                # Queue for async processing after transaction commits
+                entity_ids_for_async = list(set(link.entity_id for link in entity_links)) if entity_links else []
+                log_buffer.append(
+                    f"[11] Observations: queued {len(entity_ids_for_async)} entities for async processing"
+                )
+            else:
+                # Run synchronously inside transaction for atomicity
+                await observation_regeneration.regenerate_observations_batch(
+                    conn, embeddings_model, llm_config, bank_id, entity_links, log_buffer
+                )
+                entity_ids_for_async = []
+
            # Map results back to original content items
            result_unit_ids = _map_results_to_contents(contents, extracted_facts, is_duplicate_flags, unit_ids)

+        # Trigger background tasks AFTER transaction commits
+        await _trigger_background_tasks(task_backend, bank_id, unit_ids, non_duplicate_facts, entity_ids_for_async)
+
        # Log final summary
        total_time = time.time() - start_time
        log_buffer.append(f"{'=' * 60}")
@@ -448,3 +458,35 @@ def _map_results_to_contents(
        result_unit_ids.append(content_unit_ids)

    return result_unit_ids
+
+
+async def _trigger_background_tasks(
+    task_backend,
+    bank_id: str,
+    unit_ids: list[str],
+    facts: list[ProcessedFact],
+    entity_ids_for_observations: list[str] | None = None,
+) -> None:
+    """Trigger background tasks after transaction commits."""
+    # Trigger opinion reinforcement if there are entities
+    fact_entities = [[e.name for e in fact.entities] for fact in facts]
+    if any(fact_entities):
+        await task_backend.submit_task(
+            {
+                "type": "reinforce_opinion",
+                "bank_id": bank_id,
+                "created_unit_ids": unit_ids,
+                "unit_texts": [fact.fact_text for fact in facts],
+                "unit_entities": fact_entities,
+            }
+        )
+
+    # Trigger observation regeneration if async mode is enabled
+    if entity_ids_for_observations:
+        await task_backend.submit_task(
+            {
+                "type": "regenerate_observations",
+                "bank_id": bank_id,
+                "entity_ids": entity_ids_for_observations,
+            }
+        )
@@ -21,7 +21,6 @@ class RetainContentDict(TypedDict, total=False):
        metadata: Custom key-value metadata (optional)
        document_id: Document ID for this content item (optional)
        entities: User-provided entities to merge with extracted entities (optional)
-        tags: Visibility scope tags for this content item (optional)
    """

    content: str  # Required
@@ -30,7 +29,6 @@ class RetainContentDict(TypedDict, total=False):
    metadata: dict[str, str]
    document_id: str
    entities: list[dict[str, str]]  # [{"text": "...", "type": "..."}]
-    tags: list[str]  # Visibility scope tags


 def _now_utc() -> datetime:
@@ -51,7 +49,6 @@ class RetainContent:
    event_date: datetime = field(default_factory=_now_utc)
    metadata: dict[str, str] = field(default_factory=dict)
    entities: list[dict[str, str]] = field(default_factory=list)  # User-provided entities
-    tags: list[str] = field(default_factory=list)  # Visibility scope tags


@dataclass
@@ -86,10 +83,10 @@ class CausalRelation:
    """
    Causal relationship between facts.

-    Represents how one fact was caused by another.
+    Represents how one fact causes, enables, or prevents another.
    """

-    relation_type: str  # "caused_by"
+    relation_type: str  # "causes", "enables", "prevents", "caused_by"
    target_fact_index: int  # Index of the target fact in the batch
    strength: float = 1.0  # Strength of the causal relationship

@@ -116,7 +113,6 @@ class ExtractedFact:
    context: str = ""
    mentioned_at: datetime | None = None
    metadata: dict[str, str] = field(default_factory=dict)
-    tags: list[str] = field(default_factory=list)  # Visibility scope tags


@dataclass
@@ -162,9 +158,6 @@ class ProcessedFact:
    # Track which content this fact came from (for user entity merging)
    content_index: int = 0

-    # Visibility scope tags
-    tags: list[str] = field(default_factory=list)
-
    @property
    def is_duplicate(self) -> bool:
        """Check if this fact was marked as a duplicate."""
@@ -208,7 +201,6 @@ class ProcessedFact:
            causal_relations=extracted_fact.causal_relations,
            chunk_id=chunk_id,
            content_index=extracted_fact.content_index,
-            tags=extracted_fact.tags,
        )


@@ -240,7 +232,6 @@ class RetainBatch:
    document_id: str | None = None
    fact_type_override: str | None = None
    confidence_score: float | None = None
-    document_tags: list[str] = field(default_factory=list)  # Tags applied to all items

    # Extracted data (populated during processing)
    extracted_facts: list[ExtractedFact] = field(default_factory=list)
@@ -11,7 +11,6 @@ from abc import ABC, abstractmethod

 from ..db_utils import acquire_with_retry
 from ..memory_engine import fq_table
-from .tags import TagsMatch, filter_results_by_tags
 from .types import MPFPTimings, RetrievalResult

 logger = logging.getLogger(__name__)
@@ -44,8 +43,6 @@ class GraphRetriever(ABC):
        semantic_seeds: list[RetrievalResult] | None = None,
        temporal_seeds: list[RetrievalResult] | None = None,
        adjacency=None,  # TypedAdjacency, optional pre-loaded graph
-        tags: list[str] | None = None,  # Visibility scope tags for filtering
-        tags_match: TagsMatch = "any",  # How to match tags: 'any' (OR) or 'all' (AND)
    ) -> tuple[list[RetrievalResult], MPFPTimings | None]:
        """
        Retrieve relevant facts via graph traversal.
@@ -60,7 +57,6 @@ class GraphRetriever(ABC):
            semantic_seeds: Pre-computed semantic entry points (from semantic retrieval)
            temporal_seeds: Pre-computed temporal entry points (from temporal retrieval)
            adjacency: Pre-loaded typed adjacency graph (optional, for MPFP)
-            tags: Optional list of tags for visibility filtering (OR matching)

        Returns:
            Tuple of (List of RetrievalResult with activation scores, optional timing info)
@@ -118,8 +114,6 @@ class BFSGraphRetriever(GraphRetriever):
        semantic_seeds: list[RetrievalResult] | None = None,
        temporal_seeds: list[RetrievalResult] | None = None,
        adjacency=None,  # Not used by BFS
-        tags: list[str] | None = None,
-        tags_match: TagsMatch = "any",
    ) -> tuple[list[RetrievalResult], MPFPTimings | None]:
        """
        Retrieve facts using BFS spreading activation.
@@ -135,9 +129,7 @@ class BFSGraphRetriever(GraphRetriever):
        for interface compatibility but not used.
        """
        async with acquire_with_retry(pool) as conn:
-            results = await self._retrieve_with_conn(
-                conn, query_embedding_str, bank_id, fact_type, budget, tags=tags, tags_match=tags_match
-            )
+            results = await self._retrieve_with_conn(conn, query_embedding_str, bank_id, fact_type, budget)
            return results, None

    async def _retrieve_with_conn(
@@ -147,46 +139,33 @@ class BFSGraphRetriever(GraphRetriever):
        bank_id: str,
        fact_type: str,
        budget: int,
-        tags: list[str] | None = None,
-        tags_match: TagsMatch = "any",
    ) -> list[RetrievalResult]:
        """Internal implementation with connection."""
-        from .tags import build_tags_where_clause_simple
-
-        tags_clause = build_tags_where_clause_simple(tags, 6, match=tags_match)
-        params = [query_embedding_str, bank_id, fact_type, self.entry_point_threshold, self.entry_point_limit]
-        if tags:
-            params.append(tags)

        # Step 1: Find entry points
        entry_points = await conn.fetch(
            f"""
            SELECT id, text, context, event_date, occurred_start, occurred_end,
-                   mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+                   mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
                   1 - (embedding <=> $1::vector) AS similarity
            FROM {fq_table("memory_units")}
            WHERE bank_id = $2
              AND embedding IS NOT NULL
              AND fact_type = $3
              AND (1 - (embedding <=> $1::vector)) >= $4
-              {tags_clause}
            ORDER BY embedding <=> $1::vector
            LIMIT $5
            """,
-            *params,
+            query_embedding_str,
+            bank_id,
+            fact_type,
+            self.entry_point_threshold,
+            self.entry_point_limit,
        )

        if not entry_points:
-            logger.debug(
-                f"[BFS] No entry points found for fact_type={fact_type} (tags={tags}, tags_match={tags_match})"
-            )
            return []

-        logger.debug(
-            f"[BFS] Found {len(entry_points)} entry points for fact_type={fact_type} "
-            f"(tags={tags}, tags_match={tags_match})"
-        )
-
        # Step 2: BFS spreading activation
        visited = set()
        results = []
@@ -216,8 +195,8 @@ class BFSGraphRetriever(GraphRetriever):
                neighbors = await conn.fetch(
                    f"""
                    SELECT mu.id, mu.text, mu.context, mu.occurred_start, mu.occurred_end,
-                           mu.mentioned_at, mu.embedding, mu.fact_type,
-                           mu.document_id, mu.chunk_id, mu.tags,
+                           mu.mentioned_at, mu.access_count, mu.embedding, mu.fact_type,
+                           mu.document_id, mu.chunk_id,
                           ml.weight, ml.link_type, ml.from_unit_id
                    FROM {fq_table("memory_links")} ml
                    JOIN {fq_table("memory_units")} mu ON ml.to_unit_id = mu.id
@@ -257,8 +236,4 @@ class BFSGraphRetriever(GraphRetriever):
                            neighbor_result = RetrievalResult.from_db_row(dict(n))
                            queue.append((neighbor_result, new_activation))

-        # Apply tags filtering (BFS may traverse into memories that don't match tags criteria)
-        if tags:
-            results = filter_results_by_tags(results, tags, match=tags_match)
-
        return results
@@ -18,7 +18,6 @@ import time
 from ..db_utils import acquire_with_retry
 from ..memory_engine import fq_table
 from .graph_retrieval import GraphRetriever
-from .tags import TagsMatch, filter_results_by_tags
 from .types import MPFPTimings, RetrievalResult

 logger = logging.getLogger(__name__)
@@ -31,32 +30,26 @@ async def _find_semantic_seeds(
    fact_type: str,
    limit: int = 20,
    threshold: float = 0.3,
-    tags: list[str] | None = None,
-    tags_match: TagsMatch = "any",
 ) -> list[RetrievalResult]:
    """Find semantic seeds via embedding search."""
-    from .tags import build_tags_where_clause_simple
-
-    tags_clause = build_tags_where_clause_simple(tags, 6, match=tags_match)
-    params = [query_embedding_str, bank_id, fact_type, threshold, limit]
-    if tags:
-        params.append(tags)
-
    rows = await conn.fetch(
        f"""
        SELECT id, text, context, event_date, occurred_start, occurred_end,
-               mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+               mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
               1 - (embedding <=> $1::vector) AS similarity
        FROM {fq_table("memory_units")}
        WHERE bank_id = $2
          AND embedding IS NOT NULL
          AND fact_type = $3
          AND (1 - (embedding <=> $1::vector)) >= $4
-          {tags_clause}
        ORDER BY embedding <=> $1::vector
        LIMIT $5
        """,
-        *params,
+        query_embedding_str,
+        bank_id,
+        fact_type,
+        threshold,
+        limit,
    )
    return [RetrievalResult.from_db_row(dict(r)) for r in rows]

@@ -102,8 +95,6 @@ class LinkExpansionRetriever(GraphRetriever):
        semantic_seeds: list[RetrievalResult] | None = None,
        temporal_seeds: list[RetrievalResult] | None = None,
        adjacency=None,
-        tags: list[str] | None = None,
-        tags_match: TagsMatch = "any",
    ) -> tuple[list[RetrievalResult], MPFPTimings | None]:
        """
        Retrieve facts by expanding links from seeds.
@@ -118,7 +109,6 @@ class LinkExpansionRetriever(GraphRetriever):
            semantic_seeds: Pre-computed semantic entry points
            temporal_seeds: Pre-computed temporal entry points
            adjacency: Unused, kept for interface compatibility
-            tags: Optional list of tags for visibility filtering (OR matching)

        Returns:
            Tuple of (results, timings)
@@ -135,27 +125,15 @@ class LinkExpansionRetriever(GraphRetriever):
            else:
                seeds_start = time.time()
                all_seeds = await _find_semantic_seeds(
-                    conn,
-                    query_embedding_str,
-                    bank_id,
-                    fact_type,
-                    limit=20,
-                    threshold=0.3,
-                    tags=tags,
-                    tags_match=tags_match,
+                    conn, query_embedding_str, bank_id, fact_type, limit=20, threshold=0.3
                )
                timings.seeds_time = time.time() - seeds_start
-                logger.debug(
-                    f"[LinkExpansion] Found {len(all_seeds)} semantic seeds for fact_type={fact_type} "
-                    f"(tags={tags}, tags_match={tags_match})"
-                )

            # Add temporal seeds if provided
            if temporal_seeds:
                all_seeds.extend(temporal_seeds)

            if not all_seeds:
-                logger.info("[LinkExpansion] No seeds found, returning empty results")
                return [], timings

            seed_ids = list({s.id for s in all_seeds})
@@ -164,109 +142,37 @@ class LinkExpansionRetriever(GraphRetriever):
            # Run entity and causal expansion sequentially on same connection
            query_start = time.time()

-            # For observations, traverse through source_memory_ids to find entity connections.
-            # Observations don't have direct unit_entities - they inherit entities via their
-            # source world/experience facts.
-            #
-            # Path: observation → source_memory_ids → world fact → entities →
-            #       ALL world facts with those entities → their observations (excluding seeds)
-            if fact_type == "observation":
-                # Debug: Check what source_memory_ids exist on seed observations
-                debug_sources = await conn.fetch(
-                    f"""
-                    SELECT id, source_memory_ids
-                    FROM {fq_table("memory_units")}
-                    WHERE id = ANY($1::uuid[])
-                    """,
-                    seed_ids,
-                )
-                source_ids_found = []
-                for row in debug_sources:
-                    if row["source_memory_ids"]:
-                        source_ids_found.extend(row["source_memory_ids"])
-                logger.debug(
-                    f"[LinkExpansion] observation graph: {len(seed_ids)} seeds, "
-                    f"{len(source_ids_found)} source_memory_ids found"
-                )
-
-                entity_rows = await conn.fetch(
-                    f"""
-                    WITH seed_sources AS (
-                        -- Get source memory IDs from seed observations
-                        SELECT DISTINCT unnest(source_memory_ids) AS source_id
-                        FROM {fq_table("memory_units")}
-                        WHERE id = ANY($1::uuid[])
-                          AND source_memory_ids IS NOT NULL
-                    ),
-                    source_entities AS (
-                        -- Get entities from those source memories (filtered by frequency)
-                        SELECT DISTINCT ue.entity_id
-                        FROM seed_sources ss
-                        JOIN {fq_table("unit_entities")} ue ON ss.source_id = ue.unit_id
-                        JOIN {fq_table("entities")} e ON ue.entity_id = e.id
-                        WHERE e.mention_count < $2
-                    ),
-                    all_connected_sources AS (
-                        -- Find ALL world facts sharing those entities (don't exclude seed sources)
-                        -- The exclusion happens at the observation level, not the source level
-                        SELECT DISTINCT other_ue.unit_id AS source_id
-                        FROM source_entities se
-                        JOIN {fq_table("unit_entities")} other_ue ON se.entity_id = other_ue.entity_id
-                    )
-                    -- Find observations derived from connected source memories
-                    -- Only exclude the actual seed observations
-                    SELECT
-                        mu.id, mu.text, mu.context, mu.event_date, mu.occurred_start,
-                        mu.occurred_end, mu.mentioned_at, mu.embedding,
-                        mu.fact_type, mu.document_id, mu.chunk_id, mu.tags,
-                        COUNT(DISTINCT cs.source_id)::float AS score
-                    FROM all_connected_sources cs
-                    JOIN {fq_table("memory_units")} mu
-                        ON mu.source_memory_ids @> ARRAY[cs.source_id]
-                    WHERE mu.fact_type = 'observation'
-                      AND mu.id != ALL($1::uuid[])
-                    GROUP BY mu.id
-                    ORDER BY score DESC
-                    LIMIT $3
-                    """,
-                    seed_ids,
-                    self.max_entity_frequency,
-                    budget,
-                )
-                logger.debug(f"[LinkExpansion] observation graph: found {len(entity_rows)} connected observations")
-            else:
-                # For world/experience facts, use direct entity lookup
-                entity_rows = await conn.fetch(
-                    f"""
-                    SELECT
-                        mu.id, mu.text, mu.context, mu.event_date, mu.occurred_start,
-                        mu.occurred_end, mu.mentioned_at, mu.embedding,
-                        mu.fact_type, mu.document_id, mu.chunk_id, mu.tags,
-                        COUNT(*)::float AS score
-                    FROM {fq_table("unit_entities")} seed_ue
-                    JOIN {fq_table("entities")} e ON seed_ue.entity_id = e.id
-                    JOIN {fq_table("unit_entities")} other_ue ON seed_ue.entity_id = other_ue.entity_id
-                    JOIN {fq_table("memory_units")} mu ON other_ue.unit_id = mu.id
-                    WHERE seed_ue.unit_id = ANY($1::uuid[])
-                      AND e.mention_count < $2
-                      AND mu.id != ALL($1::uuid[])
-                      AND mu.fact_type = $3
-                    GROUP BY mu.id
-                    ORDER BY score DESC
-                    LIMIT $4
-                    """,
-                    seed_ids,
-                    self.max_entity_frequency,
-                    fact_type,
-                    budget,
-                )
+            entity_rows = await conn.fetch(
+                f"""
+                SELECT
+                    mu.id, mu.text, mu.context, mu.event_date, mu.occurred_start,
+                    mu.occurred_end, mu.mentioned_at, mu.access_count, mu.embedding,
+                    mu.fact_type, mu.document_id, mu.chunk_id,
+                    COUNT(*)::float AS score
+                FROM {fq_table("unit_entities")} seed_ue
+                JOIN {fq_table("entities")} e ON seed_ue.entity_id = e.id
+                JOIN {fq_table("unit_entities")} other_ue ON seed_ue.entity_id = other_ue.entity_id
+                JOIN {fq_table("memory_units")} mu ON other_ue.unit_id = mu.id
+                WHERE seed_ue.unit_id = ANY($1::uuid[])
+                  AND e.mention_count < $2
+                  AND mu.id != ALL($1::uuid[])
+                  AND mu.fact_type = $3
+                GROUP BY mu.id
+                ORDER BY score DESC
+                LIMIT $4
+                """,
+                seed_ids,
+                self.max_entity_frequency,
+                fact_type,
+                budget,
+            )

            causal_rows = await conn.fetch(
                f"""
                SELECT DISTINCT ON (mu.id)
                    mu.id, mu.text, mu.context, mu.event_date, mu.occurred_start,
-                    mu.occurred_end, mu.mentioned_at, mu.embedding,
-                    mu.fact_type, mu.document_id, mu.chunk_id, mu.tags,
+                    mu.occurred_end, mu.mentioned_at, mu.access_count, mu.embedding,
+                    mu.fact_type, mu.document_id, mu.chunk_id,
                    ml.weight + 1.0 AS score
                FROM {fq_table("memory_links")} ml
                JOIN {fq_table("memory_units")} mu ON ml.to_unit_id = mu.id
@@ -283,69 +189,11 @@ class LinkExpansionRetriever(GraphRetriever):
                budget,
            )

-            # Fallback: semantic/temporal/entity links from memory_links table
-            # These are secondary to entity links (via unit_entities) and causal links
-            # Weight is halved (0.5x) to prioritize primary link types
-            # Check both directions: seeds -> others AND others -> seeds
-            fallback_rows = await conn.fetch(
-                f"""
-                WITH outgoing AS (
-                    -- Links FROM seeds TO other facts
-                    SELECT mu.id, mu.text, mu.context, mu.event_date, mu.occurred_start,
-                           mu.occurred_end, mu.mentioned_at, mu.embedding,
-                           mu.fact_type, mu.document_id, mu.chunk_id, mu.tags,
-                           ml.weight
-                    FROM {fq_table("memory_links")} ml
-                    JOIN {fq_table("memory_units")} mu ON ml.to_unit_id = mu.id
-                    WHERE ml.from_unit_id = ANY($1::uuid[])
-                      AND ml.link_type IN ('semantic', 'temporal', 'entity')
-                      AND ml.weight >= $2
-                      AND mu.fact_type = $3
-                      AND mu.id != ALL($1::uuid[])
-                ),
-                incoming AS (
-                    -- Links FROM other facts TO seeds (reverse direction)
-                    SELECT mu.id, mu.text, mu.context, mu.event_date, mu.occurred_start,
-                           mu.occurred_end, mu.mentioned_at, mu.embedding,
-                           mu.fact_type, mu.document_id, mu.chunk_id, mu.tags,
-                           ml.weight
-                    FROM {fq_table("memory_links")} ml
-                    JOIN {fq_table("memory_units")} mu ON ml.from_unit_id = mu.id
-                    WHERE ml.to_unit_id = ANY($1::uuid[])
-                      AND ml.link_type IN ('semantic', 'temporal', 'entity')
-                      AND ml.weight >= $2
-                      AND mu.fact_type = $3
-                      AND mu.id != ALL($1::uuid[])
-                ),
-                combined AS (
-                    SELECT * FROM outgoing
-                    UNION ALL
-                    SELECT * FROM incoming
-                )
-                SELECT DISTINCT ON (id)
-                    id, text, context, event_date, occurred_start,
-                    occurred_end, mentioned_at, embedding,
-                    fact_type, document_id, chunk_id, tags,
-                    (MAX(weight) * 0.5) AS score
-                FROM combined
-                GROUP BY id, text, context, event_date, occurred_start,
-                         occurred_end, mentioned_at, embedding,
-                         fact_type, document_id, chunk_id, tags
-                ORDER BY id, score DESC
-                LIMIT $4
-                """,
-                seed_ids,
-                self.causal_weight_threshold,
-                fact_type,
-                budget,
-            )
-
            timings.edge_load_time = time.time() - query_start
-            timings.db_queries = 3
-            timings.edge_count = len(entity_rows) + len(causal_rows) + len(fallback_rows)
+            timings.db_queries = 2
+            timings.edge_count = len(entity_rows) + len(causal_rows)

        # Merge results, taking max score per fact
-        # Priority: entity links (unit_entities) > causal links > fallback links
        score_map: dict[str, float] = {}
        row_map: dict[str, dict] = {}

@@ -360,12 +208,6 @@ class LinkExpansionRetriever(GraphRetriever):
            if fact_id not in row_map:
                row_map[fact_id] = dict(row)

-        for row in fallback_rows:
-            fact_id = str(row["id"])
-            score_map[fact_id] = max(score_map.get(fact_id, 0), row["score"])
-            if fact_id not in row_map:
-                row_map[fact_id] = dict(row)
-
        # Sort by score and limit
        sorted_ids = sorted(score_map.keys(), key=lambda x: score_map[x], reverse=True)[:budget]
        rows = [row_map[fact_id] for fact_id in sorted_ids]
@@ -377,10 +219,6 @@ class LinkExpansionRetriever(GraphRetriever):
            result.activation = row["score"]
            results.append(result)

-        # Apply tags filtering (graph expansion may reach untagged memories)
-        if tags:
-            results = filter_results_by_tags(results, tags, match=tags_match)
-
        timings.result_count = len(results)
        timings.traverse = time.time() - start_time

@@ -23,7 +23,6 @@ from dataclasses import dataclass, field
 from ..db_utils import acquire_with_retry
 from ..memory_engine import fq_table
 from .graph_retrieval import GraphRetriever
-from .tags import TagsMatch
 from .types import MPFPTimings, RetrievalResult

 logger = logging.getLogger(__name__)
@@ -449,7 +448,7 @@ async def fetch_memory_units_by_ids(
        rows = await conn.fetch(
            f"""
            SELECT id, text, context, event_date, occurred_start, occurred_end,
-                   mentioned_at, embedding, fact_type, document_id, chunk_id, tags
+                   mentioned_at, access_count, embedding, fact_type, document_id, chunk_id
            FROM {fq_table("memory_units")}
            WHERE id = ANY($1::uuid[])
              AND fact_type = $2
@@ -504,8 +503,6 @@ class MPFPGraphRetriever(GraphRetriever):
        semantic_seeds: list[RetrievalResult] | None = None,
        temporal_seeds: list[RetrievalResult] | None = None,
        adjacency=None,  # Ignored - kept for interface compatibility
-        tags: list[str] | None = None,
-        tags_match: TagsMatch = "any",
    ) -> tuple[list[RetrievalResult], MPFPTimings | None]:
        """
        Retrieve facts using MPFP algorithm with lazy edge loading.
@@ -520,7 +517,6 @@ class MPFPGraphRetriever(GraphRetriever):
            semantic_seeds: Pre-computed semantic entry points
            temporal_seeds: Pre-computed temporal entry points
            adjacency: Ignored (kept for interface compatibility)
-            tags: Optional list of tags for visibility filtering (OR matching)

        Returns:
            Tuple of (List of RetrievalResult with activation scores, MPFPTimings)
@@ -536,13 +532,8 @@ class MPFPGraphRetriever(GraphRetriever):
        # If no semantic seeds provided, fall back to finding our own
        if not semantic_seed_nodes:
            seeds_start = time.time()
-            semantic_seed_nodes = await self._find_semantic_seeds(
-                pool, query_embedding_str, bank_id, fact_type, tags=tags, tags_match=tags_match
-            )
+            semantic_seed_nodes = await self._find_semantic_seeds(pool, query_embedding_str, bank_id, fact_type)
            timings.seeds_time = time.time() - seeds_start
-            logger.debug(
-                f"[MPFP] Found {len(semantic_seed_nodes)} semantic seeds for fact_type={fact_type} (tags={tags}, tags_match={tags_match})"
-            )

        # Collect all pattern jobs
        pattern_jobs = []
@@ -558,9 +549,6 @@ class MPFPGraphRetriever(GraphRetriever):
                pattern_jobs.append((temporal_seed_nodes, pattern))

        if not pattern_jobs:
-            logger.debug(
-                f"[MPFP] No pattern jobs (semantic_seeds={len(semantic_seed_nodes)}, temporal_seeds={len(temporal_seed_nodes)})"
-            )
            return [], timings

        timings.pattern_count = len(pattern_jobs)
@@ -599,7 +587,6 @@ class MPFPGraphRetriever(GraphRetriever):
        timings.fusion = time.time() - step_start

        if not fused:
-            logger.debug(f"[MPFP] No fused results after RRF fusion (pattern_count={len(pattern_results)})")
            return [], timings

        # Get top result IDs
@@ -609,13 +596,6 @@ class MPFPGraphRetriever(GraphRetriever):
        step_start = time.time()
        results = await fetch_memory_units_by_ids(pool, result_ids, fact_type)
        timings.fetch = time.time() - step_start
-
-        # Filter results by tags (graph traversal may have picked up unfiltered memories)
-        if tags:
-            from .tags import filter_results_by_tags
-
-            results = filter_results_by_tags(results, tags, match=tags_match)
-
        timings.result_count = len(results)

        # Add activation scores from fusion
@@ -654,17 +634,8 @@ class MPFPGraphRetriever(GraphRetriever):
        fact_type: str,
        limit: int = 20,
        threshold: float = 0.3,
-        tags: list[str] | None = None,
-        tags_match: TagsMatch = "any",
    ) -> list[SeedNode]:
        """Fallback: find semantic seeds via embedding search."""
-        from .tags import build_tags_where_clause_simple
-
-        tags_clause = build_tags_where_clause_simple(tags, 6, match=tags_match)
-        params = [query_embedding_str, bank_id, fact_type, threshold, limit]
-        if tags:
-            params.append(tags)
-
        async with acquire_with_retry(pool) as conn:
            rows = await conn.fetch(
                f"""
@@ -674,11 +645,14 @@ class MPFPGraphRetriever(GraphRetriever):
                  AND embedding IS NOT NULL
                  AND fact_type = $3
                  AND (1 - (embedding <=> $1::vector)) >= $4
-                  {tags_clause}
                ORDER BY embedding <=> $1::vector
                LIMIT $5
                """,
-                *params,
+                query_embedding_str,
+                bank_id,
+                fact_type,
+                threshold,
+                limit,
            )

        return [SeedNode(node_id=str(r["id"]), score=r["similarity"]) for r in rows]
@@ -0,0 +1,125 @@
+"""
+Observation utilities for generating entity observations from facts.
+
+Observations are objective facts synthesized from multiple memory facts
+about an entity, without personality influence.
+"""
+
+import logging
+
+from pydantic import BaseModel, Field
+
+from ..response_models import MemoryFact
+
+logger = logging.getLogger(__name__)
+
+
+class Observation(BaseModel):
+    """An observation about an entity."""
+
+    observation: str = Field(description="The observation text - a factual statement about the entity")
+
+
+class ObservationExtractionResponse(BaseModel):
+    """Response containing extracted observations."""
+
+    observations: list[Observation] = Field(default_factory=list, description="List of observations about the entity")
+
+
+def format_facts_for_observation_prompt(facts: list[MemoryFact]) -> str:
+    """Format facts as text for observation extraction prompt."""
+    import json
+
+    if not facts:
+        return "[]"
+    formatted = []
+    for fact in facts:
+        fact_obj = {"text": fact.text}
+
+        # Add context if available
+        if fact.context:
+            fact_obj["context"] = fact.context
+
+        # Add occurred_start if available
+        if fact.occurred_start:
+            fact_obj["occurred_at"] = fact.occurred_start
+
+        formatted.append(fact_obj)
+
+    return json.dumps(formatted, indent=2)
+
+
+def build_observation_prompt(
+    entity_name: str,
+    facts_text: str,
+) -> str:
+    """Build the observation extraction prompt for the LLM."""
+    return f"""Based on the following facts about "{entity_name}", generate a list of key observations.
+
+FACTS ABOUT {entity_name.upper()}:
+{facts_text}
+
+Your task: Synthesize the facts into clear, objective observations about {entity_name}.
+
+GUIDELINES:
+1. Each observation should be a factual statement about {entity_name}
+2. Combine related facts into single observations where appropriate
+3. Be objective - do not add opinions, judgments, or interpretations
+4. Focus on what we KNOW about {entity_name}, not what we assume
+5. Include observations about: identity, characteristics, roles, relationships, activities
+6. Write in third person (e.g., "John is..." not "I think John is...")
+7. If there are conflicting facts, note the most recent or most supported one
+
+EXAMPLES of good observations:
+- "John works at Google as a software engineer"
+- "John is detail-oriented and methodical in his approach"
+- "John collaborates frequently with Sarah on the AI project"
+- "John joined the company in 2023"
+
+EXAMPLES of bad observations (avoid these):
+- "John seems like a good person" (opinion/judgment)
+- "John probably likes his job" (assumption)
+- "I believe John is reliable" (first-person opinion)
+
+Generate 3-7 observations based on the available facts. If there are very few facts, generate fewer observations."""
+
+
+def get_observation_system_message() -> str:
+    """Get the system message for observation extraction."""
+    return "You are an objective observer synthesizing facts about an entity. Generate clear, factual observations without opinions or personality influence. Be concise and accurate."
+
+
+async def extract_observations_from_facts(llm_config, entity_name: str, facts: list[MemoryFact]) -> list[str]:
+    """
+    Extract observations from facts about an entity using LLM.
+
+    Args:
+        llm_config: LLM configuration to use
+        entity_name: Name of the entity to generate observations about
+        facts: List of facts mentioning the entity
+
+    Returns:
+        List of observation strings
+    """
+    if not facts:
+        return []
+
+    facts_text = format_facts_for_observation_prompt(facts)
+    prompt = build_observation_prompt(entity_name, facts_text)
+
+    try:
+        result = await llm_config.call(
+            messages=[
+                {"role": "system", "content": get_observation_system_message()},
+                {"role": "user", "content": prompt},
+            ],
+            response_format=ObservationExtractionResponse,
+            scope="memory_extract_observation",
+        )
+
+        observations = [op.observation for op in result.observations]
+        return observations
+
+    except Exception as e:
+        logger.warning(f"Failed to extract observations for {entity_name}: {str(e)}")
+        return []
@@ -20,7 +20,6 @@ from ..memory_engine import fq_table
 from .graph_retrieval import BFSGraphRetriever, GraphRetriever
 from .link_expansion_retrieval import LinkExpansionRetriever
 from .mpfp_retrieval import MPFPGraphRetriever
-from .tags import TagsMatch, build_tags_where_clause_simple
 from .types import MPFPTimings, RetrievalResult

 logger = logging.getLogger(__name__)
@@ -86,12 +85,7 @@ def set_default_graph_retriever(retriever: GraphRetriever) -> None:


 async def retrieve_semantic(
-    conn,
-    query_emb_str: str,
-    bank_id: str,
-    fact_type: str,
-    limit: int,
-    tags: list[str] | None = None,
+    conn, query_emb_str: str, bank_id: str, fact_type: str, limit: int
 ) -> list[RetrievalResult]:
    """
    Semantic retrieval via vector similarity.
@@ -102,44 +96,31 @@ async def retrieve_semantic(
        agent_id: bank ID
        fact_type: Fact type to filter
        limit: Maximum results to return
-        tags: Optional list of tags for visibility filtering (OR matching)

    Returns:
        List of RetrievalResult objects
    """
-    from .tags import TagsMatch, build_tags_where_clause_simple
-
-    tags_clause = build_tags_where_clause_simple(tags, 5)
-    params = [query_emb_str, bank_id, fact_type, limit]
-    if tags:
-        params.append(tags)
-
    results = await conn.fetch(
        f"""
-        SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+        SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
               1 - (embedding <=> $1::vector) AS similarity
        FROM {fq_table("memory_units")}
        WHERE bank_id = $2
          AND embedding IS NOT NULL
          AND fact_type = $3
          AND (1 - (embedding <=> $1::vector)) >= 0.3
-          {tags_clause}
        ORDER BY embedding <=> $1::vector
        LIMIT $4
        """,
-        *params,
+        query_emb_str,
+        bank_id,
+        fact_type,
+        limit,
    )
    return [RetrievalResult.from_db_row(dict(r)) for r in results]


-async def retrieve_bm25(
-    conn,
-    query_text: str,
-    bank_id: str,
-    fact_type: str,
-    limit: int,
-    tags: list[str] | None = None,
-) -> list[RetrievalResult]:
+async def retrieve_bm25(conn, query_text: str, bank_id: str, fact_type: str, limit: int) -> list[RetrievalResult]:
    """
    BM25 keyword retrieval via full-text search.

@@ -149,15 +130,12 @@ async def retrieve_bm25(
        agent_id: bank ID
        fact_type: Fact type to filter
        limit: Maximum results to return
-        tags: Optional list of tags for visibility filtering (OR matching)

    Returns:
        List of RetrievalResult objects
    """
    import re

-    from .tags import TagsMatch, build_tags_where_clause_simple
-
    # Sanitize query text: remove special characters that have meaning in tsquery
    # Keep only alphanumeric characters and spaces
    sanitized_text = re.sub(r"[^\w\s]", " ", query_text.lower())
@@ -173,24 +151,21 @@ async def retrieve_bm25(
    # This prevents empty results when some terms are missing
    query_tsquery = " | ".join(tokens)

-    tags_clause = build_tags_where_clause_simple(tags, 5)
-    params = [query_tsquery, bank_id, fact_type, limit]
-    if tags:
-        params.append(tags)
-
    results = await conn.fetch(
        f"""
-        SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+        SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
               ts_rank_cd(search_vector, to_tsquery('english', $1)) AS bm25_score
        FROM {fq_table("memory_units")}
        WHERE bank_id = $2
          AND fact_type = $3
          AND search_vector @@ to_tsquery('english', $1)
-          {tags_clause}
        ORDER BY bm25_score DESC
        LIMIT $4
        """,
-        *params,
+        query_tsquery,
+        bank_id,
+        fact_type,
+        limit,
    )
    return [RetrievalResult.from_db_row(dict(r)) for r in results]

@@ -202,8 +177,6 @@ async def retrieve_semantic_bm25_combined(
    bank_id: str,
    fact_types: list[str],
    limit: int,
-    tags: list[str] | None = None,
-    tags_match: TagsMatch = "any",
 ) -> dict[str, tuple[list[RetrievalResult], list[RetrievalResult]]]:
    """
    Combined semantic + BM25 retrieval for multiple fact types in a single query.
@@ -230,14 +203,10 @@ async def retrieve_semantic_bm25_combined(

    # If no valid tokens for BM25, just run semantic
    if not tokens:
-        tags_clause = build_tags_where_clause_simple(tags, 5, match=tags_match)
-        params = [query_emb_str, bank_id, fact_types, limit]
-        if tags:
-            params.append(tags)
        results = await conn.fetch(
            f"""
            WITH semantic_ranked AS (
-                SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+                SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
                       1 - (embedding <=> $1::vector) AS similarity,
                       NULL::float AS bm25_score,
                       'semantic' AS source,
@@ -247,14 +216,16 @@ async def retrieve_semantic_bm25_combined(
                  AND embedding IS NOT NULL
                  AND fact_type = ANY($3)
                  AND (1 - (embedding <=> $1::vector)) >= 0.3
-                  {tags_clause}
            )
-            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
                   similarity, bm25_score, source
            FROM semantic_ranked
            WHERE rn <= $4
            """,
-            *params,
+            query_emb_str,
+            bank_id,
+            fact_types,
+            limit,
        )
        # Group by fact_type
        result_dict: dict[str, tuple[list[RetrievalResult], list[RetrievalResult]]] = {
@@ -270,18 +241,12 @@ async def retrieve_semantic_bm25_combined(

    query_tsquery = " | ".join(tokens)

-    # Build tags clause - param 6 if tags provided
-    tags_clause = build_tags_where_clause_simple(tags, 6, match=tags_match)
-    params = [query_emb_str, bank_id, fact_types, limit, query_tsquery]
-    if tags:
-        params.append(tags)
-
    # Combined CTE query for both semantic and BM25 across all fact types
    # Uses window functions to limit per fact_type per method
    results = await conn.fetch(
        f"""
        WITH semantic_ranked AS (
-            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
                   1 - (embedding <=> $1::vector) AS similarity,
                   NULL::float AS bm25_score,
                   'semantic' AS source,
@@ -291,10 +256,9 @@ async def retrieve_semantic_bm25_combined(
              AND embedding IS NOT NULL
              AND fact_type = ANY($3)
              AND (1 - (embedding <=> $1::vector)) >= 0.3
-              {tags_clause}
        ),
        bm25_ranked AS (
-            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
                   NULL::float AS similarity,
                   ts_rank_cd(search_vector, to_tsquery('english', $5)) AS bm25_score,
                   'bm25' AS source,
@@ -303,15 +267,14 @@ async def retrieve_semantic_bm25_combined(
            WHERE bank_id = $2
              AND fact_type = ANY($3)
              AND search_vector @@ to_tsquery('english', $5)
-              {tags_clause}
        ),
        semantic AS (
-            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
                   similarity, bm25_score, source
            FROM semantic_ranked WHERE rn <= $4
        ),
        bm25 AS (
-            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
                   similarity, bm25_score, source
            FROM bm25_ranked WHERE rn <= $4
        )
@@ -319,7 +282,11 @@ async def retrieve_semantic_bm25_combined(
        UNION ALL
        SELECT * FROM bm25
        """,
-        *params,
+        query_emb_str,
+        bank_id,
+        fact_types,
+        limit,
+        query_tsquery,
    )

    # Group results by fact_type and source
@@ -346,8 +313,6 @@ async def retrieve_temporal_combined(
    end_date: datetime,
    budget: int,
    semantic_threshold: float = 0.1,
-    tags: list[str] | None = None,
-    tags_match: TagsMatch = "any",
 ) -> dict[str, list[RetrievalResult]]:
    """
    Temporal retrieval for multiple fact types in a single query.
@@ -376,17 +341,11 @@ async def retrieve_temporal_combined(
    if end_date.tzinfo is None:
        end_date = end_date.replace(tzinfo=UTC)

-    # Build tags clause
-    tags_clause = build_tags_where_clause_simple(tags, 7, match=tags_match)
-    params = [query_emb_str, bank_id, fact_types, start_date, end_date, semantic_threshold]
-    if tags:
-        params.append(tags)
-
    # Batch query: Get entry points for ALL fact types at once with window function
    entry_points = await conn.fetch(
        f"""
        WITH ranked_entries AS (
-            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+            SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
                   1 - (embedding <=> $1::vector) AS similarity,
                   ROW_NUMBER() OVER (PARTITION BY fact_type ORDER BY COALESCE(occurred_start, mentioned_at, occurred_end) DESC, embedding <=> $1::vector) AS rn
            FROM {fq_table("memory_units")}
@@ -404,13 +363,17 @@ async def retrieve_temporal_combined(
                  (occurred_end IS NOT NULL AND occurred_end BETWEEN $4 AND $5)
              )
              AND (1 - (embedding <=> $1::vector)) >= $6
-              {tags_clause}
        )
-        SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, embedding, fact_type, document_id, chunk_id, tags, similarity
+        SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, access_count, embedding, fact_type, document_id, chunk_id, similarity
        FROM ranked_entries
        WHERE rn <= 10
        """,
-        *params,
+        query_emb_str,
+        bank_id,
+        fact_types,
+        start_date,
+        end_date,
+        semantic_threshold,
    )

    if not entry_points:
@@ -473,20 +436,13 @@ async def retrieve_temporal_combined(
        budget_remaining = budget - len(ft_entry_points)
        batch_size = 20

-        # Build tags clause for spreading (use param 6 since 1-5 are used)
-        spreading_tags_clause = build_tags_where_clause_simple(tags, 6, table_alias="mu.", match=tags_match)
-
        while frontier and budget_remaining > 0:
            batch_ids = frontier[:batch_size]
            frontier = frontier[batch_size:]

-            spreading_params = [query_emb_str, batch_ids, ft, semantic_threshold, batch_size * 10]
-            if tags:
-                spreading_params.append(tags)
-
            neighbors = await conn.fetch(
                f"""
-                SELECT mu.id, mu.text, mu.context, mu.event_date, mu.occurred_start, mu.occurred_end, mu.mentioned_at, mu.embedding, mu.fact_type, mu.document_id, mu.chunk_id, mu.tags,
+                SELECT mu.id, mu.text, mu.context, mu.event_date, mu.occurred_start, mu.occurred_end, mu.mentioned_at, mu.access_count, mu.embedding, mu.fact_type, mu.document_id, mu.chunk_id,
                       ml.weight, ml.link_type, ml.from_unit_id,
                       1 - (mu.embedding <=> $1::vector) AS similarity
                FROM {fq_table("memory_links")} ml
@@ -497,11 +453,14 @@ async def retrieve_temporal_combined(
                  AND mu.fact_type = $3
                  AND mu.embedding IS NOT NULL
                  AND (1 - (mu.embedding <=> $1::vector)) >= $4
-                  {spreading_tags_clause}
                ORDER BY ml.weight DESC
                LIMIT $5
                """,
-                *spreading_params,
+                query_emb_str,
+                batch_ids,
+                ft,
+                semantic_threshold,
+                batch_size * 10,
            )

            for n in neighbors:
@@ -570,7 +529,6 @@ async def retrieve_temporal(
    end_date: datetime,
    budget: int,
    semantic_threshold: float = 0.1,
-    tags: list[str] | None = None,
 ) -> list[RetrievalResult]:
    """
    Temporal retrieval with spreading activation.
@@ -589,7 +547,6 @@ async def retrieve_temporal(
        end_date: End of time range
        budget: Node budget for spreading
        semantic_threshold: Minimum semantic similarity to include
-        tags: Optional list of tags for visibility filtering (OR matching)

    Returns:
        List of RetrievalResult objects with temporal scores
@@ -601,16 +558,9 @@ async def retrieve_temporal(
    if end_date.tzinfo is None:
        end_date = end_date.replace(tzinfo=UTC)

-    from .tags import TagsMatch, build_tags_where_clause_simple
-
-    tags_clause = build_tags_where_clause_simple(tags, 7)
-    params = [query_emb_str, bank_id, fact_type, start_date, end_date, semantic_threshold]
-    if tags:
-        params.append(tags)
-
    entry_points = await conn.fetch(
        f"""
-        SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, embedding, fact_type, document_id, chunk_id, tags,
+        SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at, access_count, embedding, fact_type, document_id, chunk_id,
               1 - (embedding <=> $1::vector) AS similarity
        FROM {fq_table("memory_units")}
        WHERE bank_id = $2
@@ -630,11 +580,15 @@ async def retrieve_temporal(
              (occurred_end IS NOT NULL AND occurred_end BETWEEN $4 AND $5)
          )
          AND (1 - (embedding <=> $1::vector)) >= $6
-          {tags_clause}
        ORDER BY COALESCE(occurred_start, mentioned_at, occurred_end) DESC, (embedding <=> $1::vector) ASC
        LIMIT 10
        """,
-        *params,
+        query_emb_str,
+        bank_id,
+        fact_type,
+        start_date,
+        end_date,
+        semantic_threshold,
    )

    if not entry_points:
@@ -691,7 +645,7 @@ async def retrieve_temporal(
        # Batch fetch all neighbors for this batch of nodes
        neighbors = await conn.fetch(
            f"""
-            SELECT mu.id, mu.text, mu.context, mu.event_date, mu.occurred_start, mu.occurred_end, mu.mentioned_at, mu.embedding, mu.fact_type, mu.document_id, mu.chunk_id,
+            SELECT mu.id, mu.text, mu.context, mu.event_date, mu.occurred_start, mu.occurred_end, mu.mentioned_at, mu.access_count, mu.embedding, mu.fact_type, mu.document_id, mu.chunk_id,
                   ml.weight, ml.link_type, ml.from_unit_id,
                   1 - (mu.embedding <=> $1::vector) AS similarity
            FROM {fq_table("memory_links")} ml
@@ -786,7 +740,6 @@ async def retrieve_parallel(
    query_analyzer: Optional["QueryAnalyzer"] = None,
    graph_retriever: GraphRetriever | None = None,
    temporal_constraint: tuple | None = None,  # Pre-extracted temporal constraint
-    tags: list[str] | None = None,  # Visibility scope tags for filtering
 ) -> ParallelRetrievalResult:
    """
    Run 3-way or 4-way parallel retrieval (adds temporal if detected).
@@ -802,7 +755,6 @@ async def retrieve_parallel(
        query_analyzer: Query analyzer to use (defaults to TransformerQueryAnalyzer)
        graph_retriever: Graph retrieval strategy (defaults to configured retriever)
        temporal_constraint: Pre-extracted temporal constraint (optional)
-        tags: Optional list of tags for visibility filtering (OR matching)

    Returns:
        ParallelRetrievalResult with semantic, bm25, graph, temporal results and timings
@@ -823,7 +775,6 @@ async def retrieve_parallel(
            retriever,
            question_date,
            query_analyzer,
-            tags=tags,
        )
    else:
        # For BFS, extract temporal constraint upfront (legacy path)
@@ -834,15 +785,7 @@ async def retrieve_parallel(
                query_text, reference_date=question_date, analyzer=query_analyzer
            )
        return await _retrieve_parallel_bfs(
-            pool,
-            query_text,
-            query_embedding_str,
-            bank_id,
-            fact_type,
-            thinking_budget,
-            temporal_constraint,
-            retriever,
-            tags=tags,
+            pool, query_text, query_embedding_str, bank_id, fact_type, thinking_budget, temporal_constraint, retriever
        )


@@ -866,7 +809,6 @@ async def _retrieve_parallel_mpfp(
    retriever: GraphRetriever,
    question_date: datetime | None = None,
    query_analyzer=None,
-    tags: list[str] | None = None,
 ) -> ParallelRetrievalResult:
    """
    MPFP retrieval with true parallelization.
@@ -888,9 +830,7 @@ async def _retrieve_parallel_mpfp(
        acquire_start = time.time()
        async with acquire_with_retry(pool) as conn:
            conn_wait = time.time() - acquire_start
-            results = await retrieve_semantic(
-                conn, query_embedding_str, bank_id, fact_type, limit=thinking_budget, tags=tags
-            )
+            results = await retrieve_semantic(conn, query_embedding_str, bank_id, fact_type, limit=thinking_budget)
        return _TimedResult(results, time.time() - start, conn_wait)

    async def run_bm25() -> _TimedResult:
@@ -899,7 +839,7 @@ async def _retrieve_parallel_mpfp(
        acquire_start = time.time()
        async with acquire_with_retry(pool) as conn:
            conn_wait = time.time() - acquire_start
-            results = await retrieve_bm25(conn, query_text, bank_id, fact_type, limit=thinking_budget, tags=tags)
+            results = await retrieve_bm25(conn, query_text, bank_id, fact_type, limit=thinking_budget)
        return _TimedResult(results, time.time() - start, conn_wait)

    async def run_graph() -> tuple[list[RetrievalResult], float, MPFPTimings | None]:
@@ -917,7 +857,6 @@ async def _retrieve_parallel_mpfp(
            query_text=query_text,
            semantic_seeds=None,  # Let MPFP find its own seeds
            temporal_seeds=None,  # Don't wait for temporal extraction
-            tags=tags,
        )
        return results, time.time() - start, mpfp_timing

@@ -1023,7 +962,7 @@ async def _get_temporal_entry_points(
    rows = await conn.fetch(
        f"""
        SELECT id, text, context, event_date, occurred_start, occurred_end, mentioned_at,
-               embedding, fact_type, document_id, chunk_id,
+               access_count, embedding, fact_type, document_id, chunk_id,
               1 - (embedding <=> $1::vector) AS similarity
        FROM {fq_table("memory_units")}
        WHERE bank_id = $2
@@ -1089,7 +1028,6 @@ async def _retrieve_parallel_bfs(
    thinking_budget: int,
    temporal_constraint: tuple | None,
    retriever: GraphRetriever,
-    tags: list[str] | None = None,
 ) -> ParallelRetrievalResult:
    """BFS retrieval: all methods run in parallel (original behavior)."""
    import time
@@ -1097,15 +1035,13 @@ async def _retrieve_parallel_bfs(
    async def run_semantic() -> _TimedResult:
        start = time.time()
        async with acquire_with_retry(pool) as conn:
-            results = await retrieve_semantic(
-                conn, query_embedding_str, bank_id, fact_type, limit=thinking_budget, tags=tags
-            )
+            results = await retrieve_semantic(conn, query_embedding_str, bank_id, fact_type, limit=thinking_budget)
        return _TimedResult(results, time.time() - start)

    async def run_bm25() -> _TimedResult:
        start = time.time()
        async with acquire_with_retry(pool) as conn:
-            results = await retrieve_bm25(conn, query_text, bank_id, fact_type, limit=thinking_budget, tags=tags)
+            results = await retrieve_bm25(conn, query_text, bank_id, fact_type, limit=thinking_budget)
        return _TimedResult(results, time.time() - start)

    async def run_graph() -> _TimedResult:
@@ -1117,7 +1053,6 @@ async def _retrieve_parallel_bfs(
            fact_type=fact_type,
            budget=thinking_budget,
            query_text=query_text,
-            tags=tags,
        )
        return _TimedResult(results, time.time() - start)

@@ -1133,7 +1068,6 @@ async def _retrieve_parallel_bfs(
                tc_end,
                budget=thinking_budget,
                semantic_threshold=0.1,
-                tags=tags,
            )
        return _TimedResult(results, time.time() - start)

@@ -1188,8 +1122,6 @@ async def retrieve_all_fact_types_parallel(
    question_date: datetime | None = None,
    query_analyzer: Optional["QueryAnalyzer"] = None,
    graph_retriever: GraphRetriever | None = None,
-    tags: list[str] | None = None,
-    tags_match: TagsMatch = "any",
 ) -> MultiFactTypeRetrievalResult:
    """
    Optimized retrieval for multiple fact types using batched queries.
@@ -1239,14 +1171,7 @@ async def retrieve_all_fact_types_parallel(

        # Semantic + BM25 combined
        semantic_bm25_results = await retrieve_semantic_bm25_combined(
-            conn,
-            query_embedding_str,
-            query_text,
-            bank_id,
-            fact_types,
-            thinking_budget,
-            tags=tags,
-            tags_match=tags_match,
+            conn, query_embedding_str, query_text, bank_id, fact_types, thinking_budget
        )
        semantic_bm25_time = time.time() - semantic_bm25_start

@@ -1263,8 +1188,6 @@ async def retrieve_all_fact_types_parallel(
                tc_end,
                budget=thinking_budget,
                semantic_threshold=0.1,
-                tags=tags,
-                tags_match=tags_match,
            )
            temporal_time = time.time() - temporal_start

@@ -1283,8 +1206,6 @@ async def retrieve_all_fact_types_parallel(
            query_text=query_text,
            semantic_seeds=None,
            temporal_seeds=None,
-            tags=tags,
-            tags_match=tags_match,
        )
        return ft, results, time.time() - graph_start, mpfp_timing

@@ -0,0 +1,159 @@
+"""
+Scoring functions for memory search and retrieval.
+
+Includes recency weighting, frequency weighting, temporal proximity,
+and similarity calculations used in memory activation and ranking.
+"""
+
+from datetime import datetime
+
+
+def cosine_similarity(vec1: list[float], vec2: list[float]) -> float:
+    """
+    Calculate cosine similarity between two vectors.
+
+    Args:
+        vec1: First vector
+        vec2: Second vector
+
+    Returns:
+        Similarity score between 0 and 1
+    """
+    if len(vec1) != len(vec2):
+        raise ValueError("Vectors must have same dimension")
+
+    dot_product = sum(a * b for a, b in zip(vec1, vec2))
+    magnitude1 = sum(a * a for a in vec1) ** 0.5
+    magnitude2 = sum(b * b for b in vec2) ** 0.5
+
+    if magnitude1 == 0 or magnitude2 == 0:
+        return 0.0
+
+    return dot_product / (magnitude1 * magnitude2)
+
+
+def calculate_recency_weight(days_since: float, half_life_days: float = 365.0) -> float:
+    """
+    Calculate recency weight using logarithmic decay.
+
+    This provides much better differentiation over long time periods compared to
+    exponential decay. Uses a log-based decay where the half-life parameter controls
+    when memories reach 50% weight.
+
+    Examples:
+        - Today (0 days): 1.0
+        - 1 year (365 days): ~0.5 (with default half_life=365)
+        - 2 years (730 days): ~0.33
+        - 5 years (1825 days): ~0.17
+        - 10 years (3650 days): ~0.09
+
+    This ensures that 2-year-old and 5-year-old memories have meaningfully
+    different weights, unlike exponential decay which makes them both ~0.
+
+    Args:
+        days_since: Number of days since the memory was created
+        half_life_days: Number of days for weight to reach 0.5 (default: 1 year)
+
+    Returns:
+        Weight between 0 and 1
+    """
+    import math
+
+    # Logarithmic decay: 1 / (1 + log(1 + days_since/half_life))
+    # This decays much slower than exponential, giving better long-term differentiation
+    normalized_age = days_since / half_life_days
+    return 1.0 / (1.0 + math.log1p(normalized_age))
+
+
+def calculate_frequency_weight(access_count: int, max_boost: float = 2.0) -> float:
+    """
+    Calculate frequency weight based on access count.
+
+    Frequently accessed memories are weighted higher.
+    Uses logarithmic scaling to avoid over-weighting.
+
+    Args:
+        access_count: Number of times the memory was accessed
+        max_boost: Maximum multiplier for frequently accessed memories
+
+    Returns:
+        Weight between 1.0 and max_boost
+    """
+    import math
+
+    if access_count <= 0:
+        return 1.0
+
+    # Logarithmic scaling: log(access_count + 1) / log(10)
+    # This gives: 0 accesses = 1.0, 9 accesses ~= 1.5, 99 accesses ~= 2.0
+    normalized = math.log(access_count + 1) / math.log(10)
+    return 1.0 + min(normalized, max_boost - 1.0)
+
+
+def calculate_temporal_anchor(occurred_start: datetime, occurred_end: datetime) -> datetime:
+    """
+    Calculate a single temporal anchor point from a temporal range.
+
+    Used for spreading activation - we need a single representative date
+    to calculate temporal proximity between facts. This simplifies the
+    range-to-range distance problem.
+
+    Strategy: Use midpoint of the range for balanced representation.
+
+    Args:
+        occurred_start: Start of temporal range
+        occurred_end: End of temporal range
+
+    Returns:
+        Single datetime representing the temporal anchor (midpoint)
+
+    Examples:
+        - Point event (July 14): start=July 14, end=July 14 → anchor=July 14
+        - Month range (February): start=Feb 1, end=Feb 28 → anchor=Feb 14
+        - Year range (2023): start=Jan 1, end=Dec 31 → anchor=July 1
+    """
+    # Calculate midpoint
+    time_delta = occurred_end - occurred_start
+    midpoint = occurred_start + (time_delta / 2)
+    return midpoint
+
+
+def calculate_temporal_proximity(anchor_a: datetime, anchor_b: datetime, half_life_days: float = 30.0) -> float:
+    """
+    Calculate temporal proximity between two temporal anchors.
+
+    Used for spreading activation to determine how "close" two facts are
+    in time. Uses logarithmic decay so that temporal similarity doesn't
+    drop off too quickly.
+
+    Args:
+        anchor_a: Temporal anchor of first fact
+        anchor_b: Temporal anchor of second fact
+        half_life_days: Number of days for proximity to reach 0.5
+                       (default: 30 days = 1 month)
+
+    Returns:
+        Proximity score in [0, 1] where:
+        - 1.0 = same day
+        - 0.5 = ~half_life days apart
+        - 0.0 = very distant in time
+
+    Examples:
+        - Same day: 1.0
+        - 1 week apart (half_life=30): ~0.7
+        - 1 month apart (half_life=30): ~0.5
+        - 1 year apart (half_life=30): ~0.2
+    """
+    import math
+
+    days_apart = abs((anchor_a - anchor_b).days)
+
+    if days_apart == 0:
+        return 1.0
+
+    # Logarithmic decay: 1 / (1 + log(1 + days_apart/half_life))
+    # Similar to calculate_recency_weight but for proximity between events
+    normalized_distance = days_apart / half_life_days
+    proximity = 1.0 / (1.0 + math.log1p(normalized_distance))
+
+    return proximity
@@ -1,172 +0,0 @@
-"""
-Tags filtering utilities for retrieval.
-
-Provides SQL building functions for filtering memories by tags.
-Supports four matching modes via TagsMatch enum:
- "any": OR matching, includes untagged memories (default, backward compatible)
- "all": AND matching, includes untagged memories
- "any_strict": OR matching, excludes untagged memories
- "all_strict": AND matching, excludes untagged memories
-
-OR matching (any/any_strict): Memory matches if ANY of its tags overlap with request tags
-AND matching (all/all_strict): Memory matches if ALL request tags are present in its tags
-"""
-
-from typing import Literal
-
-TagsMatch = Literal["any", "all", "any_strict", "all_strict"]
-
-
-def _parse_tags_match(match: TagsMatch) -> tuple[str, bool]:
-    """
-    Parse TagsMatch into operator and include_untagged flag.
-
-    Returns:
-        Tuple of (operator, include_untagged)
-        - operator: "&&" for any/any_strict, "@>" for all/all_strict
-        - include_untagged: True for any/all, False for any_strict/all_strict
-    """
-    if match == "any":
-        return "&&", True
-    elif match == "all":
-        return "@>", True
-    elif match == "any_strict":
-        return "&&", False
-    elif match == "all_strict":
-        return "@>", False
-    else:
-        # Default to "any" behavior
-        return "&&", True
-
-
-def build_tags_where_clause(
-    tags: list[str] | None,
-    param_offset: int = 1,
-    table_alias: str = "",
-    match: TagsMatch = "any",
-) -> tuple[str, list, int]:
-    """
-    Build a SQL WHERE clause for filtering by tags.
-
-    Supports four matching modes:
-    - "any" (default): OR matching, includes untagged memories
-    - "all": AND matching, includes untagged memories
-    - "any_strict": OR matching, excludes untagged memories
-    - "all_strict": AND matching, excludes untagged memories
-
-    Args:
-        tags: List of tags to filter by. If None or empty, returns empty clause (no filtering).
-        param_offset: Starting parameter number for SQL placeholders (default 1).
-        table_alias: Optional table alias prefix (e.g., "mu." for "memory_units mu").
-        match: Matching mode. Defaults to "any".
-
-    Returns:
-        Tuple of (sql_clause, params, next_param_offset):
-        - sql_clause: SQL WHERE clause string
-        - params: List of parameter values to bind
-        - next_param_offset: Next available parameter number
-
-    Example:
-        >>> clause, params, next_offset = build_tags_where_clause(['user_a'], 3, 'mu.', 'any_strict')
-        >>> print(clause)  # "AND mu.tags IS NOT NULL AND mu.tags != '{}' AND mu.tags && $3"
-    """
-    if not tags:
-        return "", [], param_offset
-
-    column = f"{table_alias}tags" if table_alias else "tags"
-    operator, include_untagged = _parse_tags_match(match)
-
-    if include_untagged:
-        # Include untagged memories (NULL or empty array) OR matching tags
-        clause = f"AND ({column} IS NULL OR {column} = '{{}}' OR {column} {operator} ${param_offset})"
-    else:
-        # Strict: only memories with matching tags (exclude NULL and empty)
-        clause = f"AND {column} IS NOT NULL AND {column} != '{{}}' AND {column} {operator} ${param_offset}"
-
-    return clause, [tags], param_offset + 1
-
-
-def build_tags_where_clause_simple(
-    tags: list[str] | None,
-    param_num: int,
-    table_alias: str = "",
-    match: TagsMatch = "any",
-) -> str:
-    """
-    Build a simple SQL WHERE clause for tags filtering.
-
-    This is a convenience version that returns just the clause string,
-    assuming the caller will add the tags array to their params list.
-
-    Args:
-        tags: List of tags to filter by. If None or empty, returns empty string.
-        param_num: Parameter number to use in the clause.
-        table_alias: Optional table alias prefix.
-        match: Matching mode. Defaults to "any".
-
-    Returns:
-        SQL clause string or empty string.
-    """
-    if not tags:
-        return ""
-
-    column = f"{table_alias}tags" if table_alias else "tags"
-    operator, include_untagged = _parse_tags_match(match)
-
-    if include_untagged:
-        # Include untagged memories (NULL or empty array) OR matching tags
-        return f"AND ({column} IS NULL OR {column} = '{{}}' OR {column} {operator} ${param_num})"
-    else:
-        # Strict: only memories with matching tags (exclude NULL and empty)
-        return f"AND {column} IS NOT NULL AND {column} != '{{}}' AND {column} {operator} ${param_num}"
-
-
-def filter_results_by_tags(
-    results: list,
-    tags: list[str] | None,
-    match: TagsMatch = "any",
-) -> list:
-    """
-    Filter retrieval results by tags in Python (for post-processing).
-
-    Used when SQL filtering isn't possible (e.g., graph traversal results).
-
-    Args:
-        results: List of RetrievalResult objects with a 'tags' attribute.
-        tags: List of tags to filter by. If None or empty, returns all results.
-        match: Matching mode. Defaults to "any".
-
-    Returns:
-        Filtered list of results.
-    """
-    if not tags:
-        return results
-
-    _, include_untagged = _parse_tags_match(match)
-    is_any_match = match in ("any", "any_strict")
-
-    tags_set = set(tags)
-    filtered = []
-
-    for result in results:
-        result_tags = getattr(result, "tags", None)
-
-        # Check if untagged
-        is_untagged = result_tags is None or len(result_tags) == 0
-
-        if is_untagged:
-            if include_untagged:
-                filtered.append(result)
-            # else: skip untagged
-        else:
-            result_tags_set = set(result_tags)
-            if is_any_match:
-                # Any overlap
-                if result_tags_set & tags_set:
-                    filtered.append(result)
-            else:
-                # All tags must be present
-                if tags_set <= result_tags_set:
-                    filtered.append(result)
-
-    return filtered
@@ -3,13 +3,31 @@ Think operation utilities for formulating answers based on agent and world facts
 """

 import logging
+import re
 from datetime import datetime

+from pydantic import BaseModel, Field
+
 from ..response_models import DispositionTraits, MemoryFact

 logger = logging.getLogger(__name__)


+class Opinion(BaseModel):
+    """An opinion formed by the bank."""
+
+    opinion: str = Field(description="The opinion or perspective with reasoning included")
+    confidence: float = Field(description="Confidence score for this opinion (0.0 to 1.0, where 1.0 is very confident)")
+
+
+class OpinionExtractionResponse(BaseModel):
+    """Response containing extracted opinions."""
+
+    opinions: list[Opinion] = Field(
+        default_factory=list, description="List of opinions formed with their supporting reasons and confidence scores"
+    )
+
+
 def describe_trait_level(value: int) -> str:
    """Convert trait value (1-5) to descriptive text."""
    levels = {1: "very low", 2: "low", 3: "moderate", 4: "high", 5: "very high"}
@@ -75,46 +93,17 @@ def format_facts_for_prompt(facts: list[MemoryFact]) -> str:
    return json.dumps(formatted, indent=2)


-def format_entity_summaries_for_prompt(entities: dict) -> str:
-    """Format entity summaries for inclusion in the reflect prompt.
-
-    Args:
-        entities: Dict mapping entity name to EntityState objects
-
-    Returns:
-        Formatted string with entity summaries, or empty string if no summaries
-    """
-    if not entities:
-        return ""
-
-    summaries = []
-    for name, state in entities.items():
-        # Get summary from observations (summary is stored as single observation)
-        if state.observations:
-            summary_text = state.observations[0].text
-            summaries.append(f"## {name}\n{summary_text}")
-
-    if not summaries:
-        return ""
-
-    return "\n\n".join(summaries)
-
-
 def build_think_prompt(
    agent_facts_text: str,
    world_facts_text: str,
+    opinion_facts_text: str,
    query: str,
    name: str,
    disposition: DispositionTraits,
    background: str,
    context: str | None = None,
-    entity_summaries_text: str | None = None,
 ) -> str:
-    """Build the think prompt for the LLM.
-
-    Note: opinion_facts_text parameter removed - opinions are now stored as mental models
-    and included via entity_summaries_text.
-    """
+    """Build the think prompt for the LLM."""
    disposition_desc = build_disposition_description(disposition)

    name_section = f"""
@@ -136,14 +125,6 @@ Your background:
 ADDITIONAL CONTEXT:
 {context}

-"""
-
-    entity_section = ""
-    if entity_summaries_text:
-        entity_section = f"""
-KEY PEOPLE, PLACES & THINGS I KNOW ABOUT:
-{entity_summaries_text}
-
 """

    return f"""Here's what I know and have experienced:
@@ -154,11 +135,14 @@ MY IDENTITY & EXPERIENCES:
 WHAT I KNOW ABOUT THE WORLD:
 {world_facts_text}

-{entity_section}{context_section}{name_section}{disposition_desc}{background_section}
+MY EXISTING OPINIONS & BELIEFS:
+{opinion_facts_text}
+
+{context_section}{name_section}{disposition_desc}{background_section}

 QUESTION: {query}

-Based on everything I know, believe, and who I am (including my name, disposition and background), here's what I genuinely think about this question. I'll draw on my experiences, knowledge, and personal traits to give you my honest perspective."""
+Based on everything I know, believe, and who I am (including my name, disposition and background), here's what I genuinely think about this question. I'll draw on my experiences, knowledge, opinions, and personal traits to give you my honest perspective."""


 def get_system_message(disposition: DispositionTraits) -> str:
@@ -191,11 +175,122 @@ def get_system_message(disposition: DispositionTraits) -> str:
    return f"You are a person with your own thoughts, experiences, opinions, and disposition. When asked a question, respond naturally like a human would - share what you genuinely think based on what you know and have experienced. {disposition_instruction} Be direct, express your views confidently, and use 'I think', 'I believe', 'in my view', etc. Respond in plain text without markdown formatting. IMPORTANT: Detect the language of the question and respond in the SAME language. Do not translate to English if the question is in another language."


+async def extract_opinions_from_text(llm_config, text: str, query: str) -> list[Opinion]:
+    """
+    Extract opinions with reasons and confidence from text using LLM.
+
+    Args:
+        llm_config: LLM configuration to use
+        text: Text to extract opinions from
+        query: The original query that prompted this response
+
+    Returns:
+        List of Opinion objects with text and confidence
+    """
+    extraction_prompt = f"""Extract any NEW opinions or perspectives from the answer below and rewrite them in FIRST-PERSON as if YOU are stating the opinion directly.
+
+ORIGINAL QUESTION:
+{query}
+
+ANSWER PROVIDED:
+{text}
+
+Your task: Find opinions in the answer and rewrite them AS IF YOU ARE THE ONE SAYING THEM.
+
+An opinion is a judgment, viewpoint, or conclusion that goes beyond just stating facts.
+
+IMPORTANT: Do NOT extract statements like:
+- "I don't have enough information"
+- "The facts don't contain information about X"
+- "I cannot answer because..."
+
+ONLY extract actual opinions about substantive topics.
+
+CRITICAL FORMAT REQUIREMENTS:
+1. **ALWAYS start with first-person phrases**: "I think...", "I believe...", "In my view...", "I've come to believe...", "Previously I thought... but now..."
+2. **NEVER use third-person**: Do NOT say "The speaker thinks..." or "They believe..." - always use "I"
+3. Include the reasoning naturally within the statement
+4. Provide a confidence score (0.0 to 1.0)
+
+CORRECT Examples (✓ FIRST-PERSON):
+- "I think Alice is more reliable because she consistently delivers on time and writes clean code"
+- "Previously I thought all engineers were equal, but now I feel that experience and track record really matter"
+- "I believe reliability is best measured by consistent output over time"
+- "I've come to believe that track records are more important than potential"
+
+WRONG Examples (✗ THIRD-PERSON - DO NOT USE):
+- "The speaker thinks Alice is more reliable"
+- "They believe reliability matters"
+- "It is believed that Alice is better"
+
+If no genuine opinions are expressed (e.g., the response just says "I don't know"), return an empty list."""
+
+    try:
+        result = await llm_config.call(
+            messages=[
+                {
+                    "role": "system",
+                    "content": "You are converting opinions from text into first-person statements. Always use 'I think', 'I believe', 'I feel', etc. NEVER use third-person like 'The speaker' or 'They'.",
+                },
+                {"role": "user", "content": extraction_prompt},
+            ],
+            response_format=OpinionExtractionResponse,
+            scope="memory_extract_opinion",
+        )
+
+        # Format opinions with confidence score and convert to first-person
+        formatted_opinions = []
+        for op in result.opinions:
+            # Convert third-person to first-person if needed
+            opinion_text = op.opinion
+
+            # Replace common third-person patterns with first-person
+            def singularize_verb(verb):
+                if verb.endswith("es"):
+                    return verb[:-1]  # believes -> believe
+                elif verb.endswith("s"):
+                    return verb[:-1]  # thinks -> think
+                return verb
+
+            # Pattern: "The speaker/user [verb]..." -> "I [verb]..."
+            match = re.match(
+                r"^(The speaker|The user|They|It is believed) (believes?|thinks?|feels?|says|asserts?|considers?)(\s+that)?(.*)$",
+                opinion_text,
+                re.IGNORECASE,
+            )
+            if match:
+                verb = singularize_verb(match.group(2))
+                that_part = match.group(3) or ""  # Keep " that" if present
+                rest = match.group(4)
+                opinion_text = f"I {verb}{that_part}{rest}"
+
+            # If still doesn't start with first-person, prepend "I believe that "
+            first_person_starters = [
+                "I think",
+                "I believe",
+                "I feel",
+                "In my view",
+                "I've come to believe",
+                "Previously I",
+            ]
+            if not any(opinion_text.startswith(starter) for starter in first_person_starters):
+                opinion_text = "I believe that " + opinion_text[0].lower() + opinion_text[1:]
+
+            formatted_opinions.append(Opinion(opinion=opinion_text, confidence=op.confidence))
+
+        return formatted_opinions
+
+    except Exception as e:
+        logger.warning(f"Failed to extract opinions: {str(e)}")
+        return []
+
+
 async def reflect(
    llm_config,
    query: str,
    experience_facts: list[str] = None,
    world_facts: list[str] = None,
+    opinion_facts: list[str] = None,
    name: str = "Assistant",
    disposition: DispositionTraits = None,
    background: str = "",
@@ -212,6 +307,7 @@ async def reflect(
        query: Question to answer
        experience_facts: List of experience/agent fact strings
        world_facts: List of world fact strings
+        opinion_facts: List of opinion fact strings
        name: Name of the agent/persona
        disposition: Disposition traits (defaults to neutral)
        background: Background information
@@ -232,15 +328,18 @@ async def reflect(

    agent_results = to_memory_facts(experience_facts or [], "experience")
    world_results = to_memory_facts(world_facts or [], "world")
+    opinion_results = to_memory_facts(opinion_facts or [], "opinion")

    # Format facts for prompt
    agent_facts_text = format_facts_for_prompt(agent_results)
    world_facts_text = format_facts_for_prompt(world_results)
+    opinion_facts_text = format_facts_for_prompt(opinion_results)

    # Build prompt
    prompt = build_think_prompt(
        agent_facts_text=agent_facts_text,
        world_facts_text=world_facts_text,
+        opinion_facts_text=opinion_facts_text,
        query=query,
        name=name,
        disposition=disposition,
@@ -11,13 +11,6 @@ from typing import Any, Literal
 from pydantic import BaseModel, Field


-class TemporalConstraint(BaseModel):
-    """Detected temporal constraint from query analysis."""
-
-    start: datetime | None = Field(default=None, description="Start of temporal range")
-    end: datetime | None = Field(default=None, description="End of temporal range")
-
-
 class QueryInfo(BaseModel):
    """Information about the search query."""

@@ -26,11 +19,6 @@ class QueryInfo(BaseModel):
    timestamp: datetime = Field(description="When the query was executed")
    budget: int = Field(description="Maximum nodes to explore")
    max_tokens: int = Field(description="Maximum tokens to return in results")
-    tags: list[str] | None = Field(default=None, description="Tags filter applied to recall")
-    tags_match: str | None = Field(default=None, description="Tags matching mode: any, all, any_strict, all_strict")
-    temporal_constraint: TemporalConstraint | None = Field(
-        default=None, description="Detected temporal range from query"
-    )


 class EntryPoint(BaseModel):
@@ -85,6 +73,7 @@ class NodeVisit(BaseModel):
    text: str = Field(description="Memory unit text content")
    context: str = Field(description="Memory unit context")
    event_date: datetime | None = Field(default=None, description="When the memory occurred")
+    access_count: int = Field(description="Number of times accessed before this search")

    # How this node was reached
    is_entry_point: bool = Field(description="Whether this is an entry point")
@@ -22,7 +22,6 @@ from .trace import (
    SearchPhaseMetrics,
    SearchSummary,
    SearchTrace,
-    TemporalConstraint,
    WeightComponents,
 )

@@ -46,14 +45,7 @@ class SearchTracer:
        json_output = trace.to_json()
    """

-    def __init__(
-        self,
-        query: str,
-        budget: int,
-        max_tokens: int,
-        tags: list[str] | None = None,
-        tags_match: str | None = None,
-    ):
+    def __init__(self, query: str, budget: int, max_tokens: int):
        """
        Initialize tracer.

@@ -61,14 +53,10 @@ class SearchTracer:
            query: Search query text
            budget: Maximum nodes to explore
            max_tokens: Maximum tokens to return in results
-            tags: Tags filter applied to recall
-            tags_match: Tags matching mode (any, all, any_strict, all_strict)
        """
        self.query_text = query
        self.budget = budget
        self.max_tokens = max_tokens
-        self.tags = tags
-        self.tags_match = tags_match

        # Trace data
        self.query_embedding: list[float] | None = None
@@ -78,9 +66,6 @@ class SearchTracer:
        self.pruned: list[PruningDecision] = []
        self.phase_metrics: list[SearchPhaseMetrics] = []

-        # Temporal constraint detected from query
-        self.temporal_constraint: TemporalConstraint | None = None
-
        # New 4-way retrieval tracking
        self.retrieval_results: list[RetrievalMethodResults] = []
        self.rrf_merged: list[RRFMergeResult] = []
@@ -103,11 +88,6 @@ class SearchTracer:
        """Record the query embedding."""
        self.query_embedding = embedding

-    def record_temporal_constraint(self, start: datetime | None, end: datetime | None):
-        """Record the detected temporal constraint from query analysis."""
-        if start is not None or end is not None:
-            self.temporal_constraint = TemporalConstraint(start=start, end=end)
-
    def add_entry_point(self, node_id: str, text: str, similarity: float, rank: int):
        """
        Record an entry point.
@@ -136,6 +116,7 @@ class SearchTracer:
        text: str,
        context: str,
        event_date: datetime | None,
+        access_count: int,
        is_entry_point: bool,
        parent_node_id: str | None,
        link_type: Literal["temporal", "semantic", "entity"] | None,
@@ -154,6 +135,7 @@ class SearchTracer:
            text: Memory unit text
            context: Memory unit context
            event_date: When the memory occurred
+            access_count: Access count before this search
            is_entry_point: Whether this is an entry point
            parent_node_id: Node that led here (None for entry points)
            link_type: Type of link from parent
@@ -192,6 +174,7 @@ class SearchTracer:
            text=text,
            context=context,
            event_date=event_date,
+            access_count=access_count,
            is_entry_point=is_entry_point,
            parent_node_id=parent_node_id,
            link_type=link_type,
@@ -330,8 +313,8 @@ class SearchTracer:
                RetrievalResult(
                    rank=rank,
                    node_id=doc_id,
-                    text=data.get("text") or "",
-                    context=data.get("context") or "",
+                    text=data.get("text", ""),
+                    context=data.get("context", ""),
                    event_date=data.get("event_date"),
                    fact_type=data.get("fact_type") or fact_type,
                    score=score,
@@ -445,9 +428,6 @@ class SearchTracer:
            timestamp=datetime.now(UTC),
            budget=self.budget,
            max_tokens=self.max_tokens,
-            tags=self.tags,
-            tags_match=self.tags_match,
-            temporal_constraint=self.temporal_constraint,
        )

        # Create summary
@@ -46,8 +46,8 @@ class RetrievalResult:
    mentioned_at: datetime | None = None
    document_id: str | None = None
    chunk_id: str | None = None
+    access_count: int = 0
    embedding: list[float] | None = None
-    tags: list[str] | None = None  # Visibility scope tags

    # Retrieval-specific scores (only one will be set depending on retrieval method)
    similarity: float | None = None  # Semantic retrieval
@@ -70,8 +70,8 @@ class RetrievalResult:
            mentioned_at=row.get("mentioned_at"),
            document_id=row.get("document_id"),
            chunk_id=row.get("chunk_id"),
+            access_count=row.get("access_count", 0),
            embedding=row.get("embedding"),
-            tags=row.get("tags"),
            similarity=row.get("similarity"),
            bm25_score=row.get("bm25_score"),
            activation=row.get("activation"),
@@ -154,8 +154,8 @@ class ScoredResult:
            "mentioned_at": self.retrieval.mentioned_at,
            "document_id": self.retrieval.document_id,
            "chunk_id": self.retrieval.chunk_id,
+            "access_count": self.retrieval.access_count,
            "embedding": self.retrieval.embedding,
-            "tags": self.retrieval.tags,
            "semantic_similarity": self.retrieval.similarity,
            "bm25_score": self.retrieval.bm25_score,
        }
@@ -1,40 +1,31 @@
 """
-Task backend for distributed task processing.
+Abstract task backend for running async tasks.

-This provides an abstraction for task storage and execution:
- BrokerTaskBackend: Uses PostgreSQL as broker (production)
- SyncTaskBackend: Executes tasks immediately (testing/embedded)
+This provides an abstraction that can be adapted to different execution models:
+- AsyncIO queue (default implementation)
+- Pub/Sub architectures (future)
+- Message brokers (future)
 """

-import json
+import asyncio
 import logging
 from abc import ABC, abstractmethod
 from collections.abc import Awaitable, Callable
-from typing import TYPE_CHECKING, Any
-
-if TYPE_CHECKING:
-    import asyncpg
+from typing import Any

 logger = logging.getLogger(__name__)


-def fq_table(table: str, schema: str | None = None) -> str:
-    """Get fully-qualified table name with optional schema prefix."""
-    if schema:
-        return f'"{schema}".{table}'
-    return table
-
-
 class TaskBackend(ABC):
    """
    Abstract base class for task execution backends.

    Implementations must:
    1. Store/publish task events (as serializable dicts)
-    2. Execute tasks through a provided executor callback (optional)
+    2. Execute tasks through a provided executor callback

    The backend treats tasks as pure dictionaries that can be serialized
-    and stored in the database. The executor (typically MemoryEngine.execute_task)
+    and sent over the network. The executor (typically MemoryEngine.execute_task)
    receives the dict and routes it to the appropriate handler.
    """

@@ -55,7 +46,7 @@ class TaskBackend(ABC):
    @abstractmethod
    async def initialize(self):
        """
-        Initialize the backend (e.g., connect to database).
+        Initialize the backend (e.g., start workers, connect to broker).
        """
        pass

@@ -72,7 +63,7 @@ class TaskBackend(ABC):
    @abstractmethod
    async def shutdown(self):
        """
-        Shutdown the backend gracefully.
+        Shutdown the backend gracefully (e.g., stop workers, close connections).
        """
        pass

@@ -102,8 +93,9 @@ class SyncTaskBackend(TaskBackend):
    """
    Synchronous task backend that executes tasks immediately.

-    This is useful for tests and embedded/CLI usage where we don't want
-    background workers. Tasks are executed inline rather than being queued.
+    This is useful for embedded/CLI usage where we don't want background
+    workers that prevent clean exit. Tasks are executed inline rather than
+    being queued.
    """

    async def initialize(self):
@@ -129,129 +121,221 @@ class SyncTaskBackend(TaskBackend):
        logger.debug("SyncTaskBackend shutdown")


-class BrokerTaskBackend(TaskBackend):
+class NoopTaskBackend(TaskBackend):
    """
-    Task backend using PostgreSQL as broker.
+    No-op task backend that discards all tasks.

-    submit_task() stores task_payload in async_operations table.
-    Actual polling and execution is handled separately by WorkerPoller.
-
-    This backend is used by the API to store tasks. Workers poll
-    the database separately to claim and execute tasks.
+    This is useful for tests where background task execution is not needed
+    and would only slow down the test suite.
    """

-    def __init__(
-        self,
-        pool_getter: Callable[[], "asyncpg.Pool"],
-        schema: str | None = None,
-        schema_getter: Callable[[], str | None] | None = None,
-    ):
-        """
-        Initialize the broker task backend.
-
-        Args:
-            pool_getter: Callable that returns the asyncpg connection pool
-            schema: Database schema for multi-tenant support (optional, static)
-            schema_getter: Callable that returns current schema dynamically (optional).
-                          If set, takes precedence over static schema for submit_task.
-        """
-        super().__init__()
-        self._pool_getter = pool_getter
-        self._schema = schema
-        self._schema_getter = schema_getter
-
    async def initialize(self):
-        """Initialize the backend."""
+        """No-op."""
        self._initialized = True
-        logger.info("BrokerTaskBackend initialized")
+        logger.debug("NoopTaskBackend initialized")
+
+    async def submit_task(self, task_dict: dict[str, Any]):
+        """Discard the task (do nothing)."""
+        pass
+
+    async def shutdown(self):
+        """No-op."""
+        self._initialized = False
+        logger.debug("NoopTaskBackend shutdown")
+
+
+class AsyncIOQueueBackend(TaskBackend):
+    """
+    Task backend implementation using asyncio queues.
+
+    This is the default implementation that uses in-process asyncio queues
+    and a periodic consumer worker.
+    """
+
+    def __init__(self, batch_size: int = 10, batch_interval: float = 1.0):
+        """
+        Initialize AsyncIO queue backend.
+
+        Args:
+            batch_size: Maximum number of tasks to process in one batch
+            batch_interval: Maximum time (seconds) to wait before processing batch
+        """
+        super().__init__()
+        self._queue: asyncio.Queue | None = None
+        self._worker_task: asyncio.Task | None = None
+        self._shutdown_event: asyncio.Event | None = None
+        self._batch_size = batch_size
+        self._batch_interval = batch_interval
+        self._in_flight_count = 0
+        self._in_flight_lock = asyncio.Lock()
+
+    async def initialize(self):
+        """Initialize the queue and start the worker."""
+        if self._initialized:
+            return
+
+        self._queue = asyncio.Queue()
+        self._shutdown_event = asyncio.Event()
+        self._worker_task = asyncio.create_task(self._worker())
+        self._initialized = True
+        logger.info("AsyncIOQueueBackend initialized")

    async def submit_task(self, task_dict: dict[str, Any]):
        """
-        Store task payload in async_operations table.
-
-        The task_dict should contain an 'operation_id' if updating an existing
-        operation record, otherwise a new operation will be created.
+        Submit a task by putting it in the queue.

        Args:
-            task_dict: Task dictionary to store (must be JSON serializable)
+            task_dict: Task dictionary to execute
        """
        if not self._initialized:
            await self.initialize()

-        pool = self._pool_getter()
-        operation_id = task_dict.get("operation_id")
-        task_type = task_dict.get("type", "unknown")
-        bank_id = task_dict.get("bank_id")
-        payload_json = json.dumps(task_dict)
-
-        schema = self._schema_getter() if self._schema_getter else self._schema
-        table = fq_table("async_operations", schema)
-
-        if operation_id:
-            # Update existing operation with task payload
-            await pool.execute(
-                f"""
-                UPDATE {table}
-                SET task_payload = $1::jsonb, updated_at = now()
-                WHERE operation_id = $2
-                """,
-                payload_json,
-                operation_id,
-            )
-            logger.debug(f"Updated task payload for operation {operation_id}")
-        else:
-            # Insert new operation (for tasks without pre-created records)
-            # e.g., access_count_update tasks
-            import uuid
-
-            new_id = uuid.uuid4()
-            await pool.execute(
-                f"""
-                INSERT INTO {table} (operation_id, bank_id, operation_type, status, task_payload)
-                VALUES ($1, $2, $3, 'pending', $4::jsonb)
-                """,
-                new_id,
-                bank_id,
-                task_type,
-                payload_json,
-            )
-            logger.debug(f"Created new operation {new_id} for task type {task_type}")
-
-    async def shutdown(self):
-        """Shutdown the backend."""
-        self._initialized = False
-        logger.info("BrokerTaskBackend shutdown")
+        await self._queue.put(task_dict)

    async def wait_for_pending_tasks(self, timeout: float = 120.0):
        """
-        Wait for pending tasks to be processed.
+        Wait for all pending tasks in the queue and in-flight tasks to complete.

-        In the broker model, this polls the database to check if tasks
-        for this process have been completed. This is useful in tests
-        when worker_enabled=True (API processes its own tasks).
+        This is useful in tests to ensure background tasks complete before assertions.

        Args:
-            timeout: Maximum time to wait in seconds
+            timeout: Maximum time to wait in seconds (default 120s for long-running tasks)
        """
-        import asyncio
-
-        pool = self._pool_getter()
-        schema = self._schema_getter() if self._schema_getter else self._schema
-        table = fq_table("async_operations", schema)
+        if not self._initialized or self._queue is None:
+            return

+        # Wait for queue to be empty AND no in-flight tasks
        start_time = asyncio.get_event_loop().time()
        while asyncio.get_event_loop().time() - start_time < timeout:
-            # Check if there are any pending tasks with payloads
-            count = await pool.fetchval(
-                f"""
-                SELECT COUNT(*) FROM {table}
-                WHERE status = 'pending' AND task_payload IS NOT NULL
-                """
-            )
+            async with self._in_flight_lock:
+                in_flight = self._in_flight_count

-            if count == 0:
+            if self._queue.empty() and in_flight == 0:
+                # Queue is empty and no tasks in flight, we're done
                return

+            # Wait a bit before checking again
            await asyncio.sleep(0.5)

-        logger.warning(f"Timeout waiting for pending tasks after {timeout}s")
+    async def shutdown(self):
+        """Shutdown the worker and drain the queue."""
+        if not self._initialized:
+            return
+
+        logger.info("Shutting down AsyncIOQueueBackend...")
+
+        # Signal shutdown
+        self._shutdown_event.set()
+
+        # Cancel worker
+        if self._worker_task is not None:
+            self._worker_task.cancel()
+            try:
+                await self._worker_task
+            except asyncio.CancelledError:
+                pass  # Worker cancelled successfully
+
+        self._initialized = False
+        logger.info("AsyncIOQueueBackend shutdown complete")
+
+    async def _execute_task_with_tracking(self, task_dict: dict[str, Any]):
+        """Execute a task and track its in-flight status."""
+        async with self._in_flight_lock:
+            self._in_flight_count += 1
+        try:
+            await self._execute_task(task_dict)
+        finally:
+            async with self._in_flight_lock:
+                self._in_flight_count -= 1
+
+    async def _execute_task_no_tracking(self, task_dict: dict[str, Any]):
+        """Execute a task without in-flight tracking (tracking done at batch level)."""
+        await self._execute_task(task_dict)
+
+    def _get_queue_stats(self) -> tuple[int, dict[str, int]]:
+        """Get current queue size and bank_id distribution."""
+        queue_size = self._queue.qsize() if self._queue else 0
+        bank_distribution: dict[str, int] = {}
+
+        if queue_size > 0 and self._queue:
+            # Peek at queue items without removing them
+            # Note: This is a snapshot and may not be perfectly accurate due to concurrency
+            try:
+                # Access internal deque for logging purposes only
+                items = list(self._queue._queue)  # type: ignore[attr-defined]
+                for item in items:
+                    bank_id = item.get("bank_id", "unknown")
+                    bank_distribution[bank_id] = bank_distribution.get(bank_id, 0) + 1
+            except Exception:
+                pass  # Queue access failed, return empty distribution
+
+        return queue_size, bank_distribution
+
+    async def _worker(self):
+        """
+        Background worker that processes tasks in batches.
+
+        Collects tasks for up to batch_interval seconds or batch_size items,
+        then processes them.
+        """
+        while not self._shutdown_event.is_set():
+            try:
+                # Collect tasks for batching
+                tasks = []
+                deadline = asyncio.get_event_loop().time() + self._batch_interval
+
+                while len(tasks) < self._batch_size and asyncio.get_event_loop().time() < deadline:
+                    try:
+                        remaining_time = max(0.1, deadline - asyncio.get_event_loop().time())
+                        task_dict = await asyncio.wait_for(self._queue.get(), timeout=remaining_time)
+                        # Track task as in-flight immediately when picked up from queue
+                        # This prevents wait_for_pending_tasks from returning too early
+                        async with self._in_flight_lock:
+                            self._in_flight_count += 1
+                        tasks.append(task_dict)
+                    except TimeoutError:
+                        break
+
+                # Process batch
+                if tasks:
+                    # Log batch start with queue stats
+                    queue_size, bank_distribution = self._get_queue_stats()
+
+                    # Summarize batch by task type and bank
+                    batch_summary: dict[str, dict[str, int]] = {}
+                    for task_dict in tasks:
+                        task_type = task_dict.get("type", "unknown")
+                        bank_id = task_dict.get("bank_id", "unknown")
+                        if task_type not in batch_summary:
+                            batch_summary[task_type] = {}
+                        batch_summary[task_type][bank_id] = batch_summary[task_type].get(bank_id, 0) + 1
+
+                    # Build log message
+                    batch_parts = []
+                    for task_type, banks in sorted(batch_summary.items()):
+                        bank_str = ", ".join(f"{b}:{c}" for b, c in sorted(banks.items()))
+                        batch_parts.append(f"{task_type}[{bank_str}]")
+                    batch_str = ", ".join(batch_parts)
+
+                    if queue_size > 0:
+                        pending_str = ", ".join(f"{k}:{v}" for k, v in sorted(bank_distribution.items()))
+                        logger.info(
+                            f"Processing {len(tasks)} tasks: {batch_str} (pending={queue_size} [{pending_str}])"
+                        )
+                    else:
+                        logger.info(f"Processing {len(tasks)} tasks: {batch_str}")
+
+                    # Execute tasks concurrently (in_flight already tracked when picked up)
+                    await asyncio.gather(
+                        *[self._execute_task_no_tracking(task_dict) for task_dict in tasks], return_exceptions=True
+                    )
+
+                    # Decrement in_flight count after all tasks complete
+                    async with self._in_flight_lock:
+                        self._in_flight_count -= len(tasks)
+
+            except asyncio.CancelledError:
+                break
+            except Exception as e:
+                logger.error(f"Worker error: {e}")
+                await asyncio.sleep(1)  # Backoff on error
@@ -65,3 +65,154 @@ async def extract_facts(
        return [], chunks

    return facts, chunks
+
+
+def cosine_similarity(vec1: list[float], vec2: list[float]) -> float:
+    """
+    Calculate cosine similarity between two vectors.
+
+    Args:
+        vec1: First vector
+        vec2: Second vector
+
+    Returns:
+        Similarity score between 0 and 1
+    """
+    if len(vec1) != len(vec2):
+        raise ValueError("Vectors must have same dimension")
+
+    dot_product = sum(a * b for a, b in zip(vec1, vec2))
+    magnitude1 = sum(a * a for a in vec1) ** 0.5
+    magnitude2 = sum(b * b for b in vec2) ** 0.5
+
+    if magnitude1 == 0 or magnitude2 == 0:
+        return 0.0
+
+    return dot_product / (magnitude1 * magnitude2)
+
+
+def calculate_recency_weight(days_since: float, half_life_days: float = 365.0) -> float:
+    """
+    Calculate recency weight using logarithmic decay.
+
+    This provides much better differentiation over long time periods compared to
+    exponential decay. Uses a log-based decay where the half-life parameter controls
+    when memories reach 50% weight.
+
+    Examples:
+        - Today (0 days): 1.0
+        - 1 year (365 days): ~0.5 (with default half_life=365)
+        - 2 years (730 days): ~0.33
+        - 5 years (1825 days): ~0.17
+        - 10 years (3650 days): ~0.09
+
+    This ensures that 2-year-old and 5-year-old memories have meaningfully
+    different weights, unlike exponential decay which makes them both ~0.
+
+    Args:
+        days_since: Number of days since the memory was created
+        half_life_days: Number of days for weight to reach 0.5 (default: 1 year)
+
+    Returns:
+        Weight between 0 and 1
+    """
+    import math
+
+    # Logarithmic decay: 1 / (1 + log(1 + days_since/half_life))
+    # This decays much slower than exponential, giving better long-term differentiation
+    normalized_age = days_since / half_life_days
+    return 1.0 / (1.0 + math.log1p(normalized_age))
+
+
+def calculate_frequency_weight(access_count: int, max_boost: float = 2.0) -> float:
+    """
+    Calculate frequency weight based on access count.
+
+    Frequently accessed memories are weighted higher.
+    Uses logarithmic scaling to avoid over-weighting.
+
+    Args:
+        access_count: Number of times the memory was accessed
+        max_boost: Maximum multiplier for frequently accessed memories
+
+    Returns:
+        Weight between 1.0 and max_boost
+    """
+    import math
+
+    if access_count <= 0:
+        return 1.0
+
+    # Logarithmic scaling: log(access_count + 1) / log(10)
+    # This gives: 0 accesses = 1.0, 9 accesses ~= 1.5, 99 accesses ~= 2.0
+    normalized = math.log(access_count + 1) / math.log(10)
+    return 1.0 + min(normalized, max_boost - 1.0)
+
+
+def calculate_temporal_anchor(occurred_start: datetime, occurred_end: datetime) -> datetime:
+    """
+    Calculate a single temporal anchor point from a temporal range.
+
+    Used for spreading activation - we need a single representative date
+    to calculate temporal proximity between facts. This simplifies the
+    range-to-range distance problem.
+
+    Strategy: Use midpoint of the range for balanced representation.
+
+    Args:
+        occurred_start: Start of temporal range
+        occurred_end: End of temporal range
+
+    Returns:
+        Single datetime representing the temporal anchor (midpoint)
+
+    Examples:
+        - Point event (July 14): start=July 14, end=July 14 → anchor=July 14
+        - Month range (February): start=Feb 1, end=Feb 28 → anchor=Feb 14
+        - Year range (2023): start=Jan 1, end=Dec 31 → anchor=July 1
+    """
+    # Calculate midpoint
+    time_delta = occurred_end - occurred_start
+    midpoint = occurred_start + (time_delta / 2)
+    return midpoint
+
+
+def calculate_temporal_proximity(anchor_a: datetime, anchor_b: datetime, half_life_days: float = 30.0) -> float:
+    """
+    Calculate temporal proximity between two temporal anchors.
+
+    Used for spreading activation to determine how "close" two facts are
+    in time. Uses logarithmic decay so that temporal similarity doesn't
+    drop off too quickly.
+
+    Args:
+        anchor_a: Temporal anchor of first fact
+        anchor_b: Temporal anchor of second fact
+        half_life_days: Number of days for proximity to reach 0.5
+                       (default: 30 days = 1 month)
+
+    Returns:
+        Proximity score in [0, 1] where:
+        - 1.0 = same day
+        - 0.5 = ~half_life days apart
+        - 0.0 = very distant in time
+
+    Examples:
+        - Same day: 1.0
+        - 1 week apart (half_life=30): ~0.7
+        - 1 month apart (half_life=30): ~0.5
+        - 1 year apart (half_life=30): ~0.2
+    """
+    import math
+
+    days_apart = abs((anchor_a - anchor_b).days)
+
+    if days_apart == 0:
+        return 1.0
+
+    # Logarithmic decay: 1 / (1 + log(1 + days_apart/half_life))
+    # Similar to calculate_recency_weight but for proximity between events
+    normalized_distance = days_apart / half_life_days
+    proximity = 1.0 / (1.0 + math.log1p(normalized_distance))
+
+    return proximity
@@ -21,10 +21,6 @@ from hindsight_api.extensions.context import DefaultExtensionContext, ExtensionC
 from hindsight_api.extensions.http import HttpExtension
 from hindsight_api.extensions.loader import load_extension
 from hindsight_api.extensions.operation_validator import (
-    # Consolidation operation
-    ConsolidateContext,
-    ConsolidateResult,
-    # Core operations
    OperationValidationError,
    OperationValidatorExtension,
    RecallContext,
@@ -37,7 +33,6 @@ from hindsight_api.extensions.operation_validator import (
 )
 from hindsight_api.extensions.tenant import (
    AuthenticationError,
-    Tenant,
    TenantContext,
    TenantExtension,
 )
@@ -52,7 +47,7 @@ __all__ = [
    "DefaultExtensionContext",
    # HTTP Extension
    "HttpExtension",
-    # Operation Validator - Core
+    # Operation Validator
    "OperationValidationError",
    "OperationValidatorExtension",
    "RecallContext",
@@ -62,14 +57,10 @@ __all__ = [
    "RetainContext",
    "RetainResult",
    "ValidationResult",
-    # Operation Validator - Consolidation
-    "ConsolidateContext",
-    "ConsolidateResult",
    # Tenant/Auth
    "ApiKeyTenantExtension",
    "AuthenticationError",
    "RequestContext",
-    "Tenant",
    "TenantContext",
    "TenantExtension",
 ]
@@ -1,6 +1,6 @@
 """Built-in tenant extension implementations."""

-from hindsight_api.extensions.tenant import AuthenticationError, Tenant, TenantContext, TenantExtension
+from hindsight_api.extensions.tenant import AuthenticationError, TenantContext, TenantExtension
 from hindsight_api.models import RequestContext


@@ -31,7 +31,3 @@ class ApiKeyTenantExtension(TenantExtension):
        if context.api_key != self.expected_api_key:
            raise AuthenticationError("Invalid API key")
        return TenantContext(schema_name="public")
-
-    async def list_tenants(self) -> list[Tenant]:
-        """Return public schema for single-tenant setup."""
-        return [Tenant(schema="public")]
@@ -1,4 +1,4 @@
-"""Operation Validator Extension for validating retain/recall/reflect/consolidate operations."""
+"""Operation Validator Extension for validating retain/recall/reflect operations."""

 from abc import ABC, abstractmethod
 from dataclasses import dataclass, field
@@ -97,19 +97,6 @@ class ReflectContext:
    context: str | None = None


-# =============================================================================
-# Consolidation Pre-operation Context
-# =============================================================================
-
-
-@dataclass
-class ConsolidateContext:
-    """Context for a consolidation operation validation (pre-operation)."""
-
-    bank_id: str
-    request_context: "RequestContext"
-
-
 # =============================================================================
 # Post-operation Contexts (includes results)
 # =============================================================================
@@ -177,28 +164,9 @@ class ReflectResultContext:
    error: str | None = None


-# =============================================================================
-# Consolidation Post-operation Context
-# =============================================================================
-
-
-@dataclass
-class ConsolidateResult:
-    """Result context for post-consolidation hook."""
-
-    bank_id: str
-    request_context: "RequestContext"
-    # Result
-    processed: int = 0
-    created: int = 0
-    updated: int = 0
-    success: bool = True
-    error: str | None = None
-
-
 class OperationValidatorExtension(Extension, ABC):
    """
-    Validates and hooks into retain/recall/reflect/consolidate operations.
+    Validates and hooks into retain/recall/reflect operations.

    This extension allows implementing custom logic such as:
    - Rate limiting (pre-operation)
@@ -217,13 +185,9 @@ class OperationValidatorExtension(Extension, ABC):
        -> config = {"max_requests": "100"}

    Hook execution order:
-        1. validate_* (pre-operation)
+        1. validate_retain/validate_recall/validate_reflect (pre-operation)
        2. [operation executes]
-        3. on_*_complete (post-operation)
-
-    Supported operations:
-        - retain, recall, reflect (core memory operations)
-        - consolidate (mental models consolidation)
+        3. on_retain_complete/on_recall_complete/on_reflect_complete (post-operation)
    """

    # =========================================================================
@@ -361,44 +325,3 @@ class OperationValidatorExtension(Extension, ABC):
                - error: Error message (if failed)
        """
        pass
-
-    # =========================================================================
-    # Consolidation - Pre-operation validation hook (optional - override to implement)
-    # =========================================================================
-
-    async def validate_consolidate(self, ctx: ConsolidateContext) -> ValidationResult:
-        """
-        Validate a consolidation operation before execution.
-
-        Override to implement custom validation logic for consolidation.
-
-        Args:
-            ctx: Context containing:
-                - bank_id: Bank identifier
-                - request_context: Request context with auth info
-
-        Returns:
-            ValidationResult indicating whether the operation is allowed.
-        """
-        return ValidationResult.accept()
-
-    # =========================================================================
-    # Consolidation - Post-operation hook (optional - override to implement)
-    # =========================================================================
-
-    async def on_consolidate_complete(self, result: ConsolidateResult) -> None:
-        """
-        Called after a consolidation operation completes (success or failure).
-
-        Override to implement post-operation logic such as usage tracking or audit logging.
-
-        Args:
-            result: Result context containing:
-                - bank_id: Bank identifier
-                - processed: Number of memories processed
-                - created: Number of mental models created
-                - updated: Number of mental models updated
-                - success: Whether the operation succeeded
-                - error: Error message (if failed)
-        """
-        pass
@@ -28,18 +28,6 @@ class TenantContext:
    schema_name: str


-@dataclass
-class Tenant:
-    """
-    Represents a tenant for worker discovery.
-
-    Used by list_tenants() to return tenant information including
-    the PostgreSQL schema name for database operations.
-    """
-
-    schema: str
-
-
 class TenantExtension(Extension, ABC):
    """
    Extension for multi-tenancy and API key authentication.
@@ -73,17 +61,3 @@ class TenantExtension(Extension, ABC):
            AuthenticationError: If authentication fails.
        """
        ...
-
-    @abstractmethod
-    async def list_tenants(self) -> list[Tenant]:
-        """
-        List all tenants that should be processed by workers.
-
-        This method is used by the worker to discover all tenants that need
-        task polling. Workers will poll for pending tasks in each tenant's schema.
-
-        Returns:
-            List of Tenant objects containing schema information.
-            For single-tenant setups, return [Tenant(schema="public")].
-        """
-        ...
@@ -184,39 +184,30 @@ def main():
            reflect_llm_api_key=config.reflect_llm_api_key,
            reflect_llm_model=config.reflect_llm_model,
            reflect_llm_base_url=config.reflect_llm_base_url,
-            consolidation_llm_provider=config.consolidation_llm_provider,
-            consolidation_llm_api_key=config.consolidation_llm_api_key,
-            consolidation_llm_model=config.consolidation_llm_model,
-            consolidation_llm_base_url=config.consolidation_llm_base_url,
            embeddings_provider=config.embeddings_provider,
            embeddings_local_model=config.embeddings_local_model,
            embeddings_tei_url=config.embeddings_tei_url,
-            embeddings_openai_base_url=config.embeddings_openai_base_url,
-            embeddings_cohere_base_url=config.embeddings_cohere_base_url,
            reranker_provider=config.reranker_provider,
            reranker_local_model=config.reranker_local_model,
            reranker_tei_url=config.reranker_tei_url,
            reranker_tei_batch_size=config.reranker_tei_batch_size,
            reranker_tei_max_concurrent=config.reranker_tei_max_concurrent,
            reranker_max_candidates=config.reranker_max_candidates,
-            reranker_cohere_base_url=config.reranker_cohere_base_url,
            host=args.host,
            port=args.port,
            log_level=args.log_level,
-            log_format=config.log_format,
            mcp_enabled=config.mcp_enabled,
            graph_retriever=config.graph_retriever,
            mpfp_top_k_neighbors=config.mpfp_top_k_neighbors,
            recall_max_concurrent=config.recall_max_concurrent,
            recall_connection_budget=config.recall_connection_budget,
+            observation_min_facts=config.observation_min_facts,
+            observation_top_entities=config.observation_top_entities,
            retain_max_completion_tokens=config.retain_max_completion_tokens,
            retain_chunk_size=config.retain_chunk_size,
            retain_extract_causal_links=config.retain_extract_causal_links,
            retain_extraction_mode=config.retain_extraction_mode,
-            retain_custom_instructions=config.retain_custom_instructions,
            retain_observations_async=config.retain_observations_async,
-            enable_observations=config.enable_observations,
-            consolidation_batch_size=config.consolidation_batch_size,
            skip_llm_verification=config.skip_llm_verification,
            lazy_reranker=config.lazy_reranker,
            run_migrations_on_startup=config.run_migrations_on_startup,
@@ -224,14 +215,9 @@ def main():
            db_pool_max_size=config.db_pool_max_size,
            db_command_timeout=config.db_command_timeout,
            db_acquire_timeout=config.db_acquire_timeout,
-            worker_enabled=config.worker_enabled,
-            worker_id=config.worker_id,
-            worker_poll_interval_ms=config.worker_poll_interval_ms,
-            worker_max_retries=config.worker_max_retries,
-            worker_batch_size=config.worker_batch_size,
-            worker_http_port=config.worker_http_port,
-            reflect_max_iterations=config.reflect_max_iterations,
-            mental_model_refresh_concurrency=config.mental_model_refresh_concurrency,
+            task_backend=config.task_backend,
+            task_backend_memory_batch_size=config.task_backend_memory_batch_size,
+            task_backend_memory_batch_interval=config.task_backend_memory_batch_interval,
        )
    config.configure_logging()
    if not args.daemon:
@@ -44,6 +44,7 @@ import os
 import sys

 from mcp.server.fastmcp import FastMCP
+from mcp.types import Icon

 from hindsight_api.config import (
    DEFAULT_MCP_LOCAL_BANK_ID,
@@ -52,7 +53,6 @@ from hindsight_api.config import (
    ENV_MCP_INSTRUCTIONS,
    ENV_MCP_LOCAL_BANK_ID,
 )
-from hindsight_api.mcp_tools import MCPToolsConfig, register_mcp_tools

 # Configure logging - default to warning to avoid polluting stderr during MCP init
 # MCP clients interpret stderr output as errors, so we suppress INFO logs by default
@@ -85,6 +85,9 @@ def create_local_mcp_server(bank_id: str, memory=None) -> FastMCP:
    """
    # Import here to avoid slow startup if just checking --help
    from hindsight_api import MemoryEngine
+    from hindsight_api.engine.memory_engine import Budget
+    from hindsight_api.engine.response_models import VALID_RECALL_FACT_TYPES
+    from hindsight_api.models import RequestContext

    # Create memory engine with pg0 embedded database if not provided
    if memory is None:
@@ -102,17 +105,55 @@ def create_local_mcp_server(bank_id: str, memory=None) -> FastMCP:

    mcp = FastMCP("hindsight")

-    # Configure and register tools using shared module
-    config = MCPToolsConfig(
-        bank_id_resolver=lambda: bank_id,
-        include_bank_id_param=False,  # Local MCP uses fixed bank_id
-        tools={"retain", "recall"},  # Local MCP only has retain and recall
-        retain_description=retain_description,
-        recall_description=recall_description,
-        retain_fire_and_forget=True,  # Local MCP uses fire-and-forget pattern
-    )
+    @mcp.tool(description=retain_description)
+    async def retain(content: str, context: str = "general") -> dict:
+        """
+        Args:
+            content: The fact/memory to store (be specific and include relevant details)
+            context: Category for the memory (e.g., 'preferences', 'work', 'hobbies', 'family'). Default: 'general'
+        """
+        import asyncio

-    register_mcp_tools(mcp, memory, config)
+        async def _retain():
+            try:
+                await memory.retain_batch_async(
+                    bank_id=bank_id,
+                    contents=[{"content": content, "context": context}],
+                    request_context=RequestContext(),
+                )
+            except Exception as e:
+                logger.error(f"Error storing memory: {e}", exc_info=True)
+
+        # Fire and forget - don't block on memory storage
+        asyncio.create_task(_retain())
+        return {"status": "accepted", "message": "Memory storage initiated"}
+
+    @mcp.tool(description=recall_description)
+    async def recall(query: str, max_tokens: int = 4096, budget: str = "low") -> dict:
+        """
+        Args:
+            query: Natural language search query (e.g., "user's food preferences", "what projects is user working on")
+            max_tokens: Maximum tokens to return in results (default: 4096)
+            budget: Search budget level - "low", "mid", or "high" (default: "low")
+        """
+        try:
+            # Map string budget to enum
+            budget_map = {"low": Budget.LOW, "mid": Budget.MID, "high": Budget.HIGH}
+            budget_enum = budget_map.get(budget.lower(), Budget.LOW)
+
+            search_result = await memory.recall_async(
+                bank_id=bank_id,
+                query=query,
+                fact_type=list(VALID_RECALL_FACT_TYPES),
+                budget=budget_enum,
+                max_tokens=max_tokens,
+                request_context=RequestContext(),
+            )
+
+            return search_result.model_dump()
+        except Exception as e:
+            logger.error(f"Error searching: {e}", exc_info=True)
+            return {"error": str(e), "results": []}

    return mcp

@@ -1,494 +0,0 @@
-"""Shared MCP tool implementations for Hindsight.
-
-This module provides the core tool logic used by both:
- mcp_local.py (stdio transport for Claude Code)
- api/mcp.py (HTTP transport for API server)
-"""
-
-import json
-import logging
-from dataclasses import dataclass
-from datetime import datetime
-from typing import Any, Callable
-
-from fastmcp import FastMCP
-
-from hindsight_api import MemoryEngine
-from hindsight_api.config import (
-    DEFAULT_MCP_RECALL_DESCRIPTION,
-    DEFAULT_MCP_RETAIN_DESCRIPTION,
-)
-from hindsight_api.engine.memory_engine import Budget
-from hindsight_api.engine.response_models import VALID_RECALL_FACT_TYPES
-from hindsight_api.models import RequestContext
-
-logger = logging.getLogger(__name__)
-
-
-@dataclass
-class MCPToolsConfig:
-    """Configuration for MCP tools registration."""
-
-    # How to resolve bank_id for operations
-    bank_id_resolver: Callable[[], str | None]
-
-    # Whether to include bank_id as a parameter on tools (for multi-bank support)
-    include_bank_id_param: bool = False
-
-    # Which tools to register
-    tools: set[str] | None = None  # None means all tools
-
-    # Custom descriptions (if None, uses defaults)
-    retain_description: str | None = None
-    recall_description: str | None = None
-
-    # Retain behavior
-    retain_fire_and_forget: bool = False  # If True, use asyncio.create_task pattern
-
-
-def parse_timestamp(timestamp: str) -> datetime | None:
-    """Parse an ISO format timestamp string.
-
-    Args:
-        timestamp: ISO format timestamp (e.g., '2024-01-15T10:30:00Z')
-
-    Returns:
-        Parsed datetime or None if invalid
-
-    Raises:
-        ValueError: If timestamp format is invalid
-    """
-    try:
-        return datetime.fromisoformat(timestamp.replace("Z", "+00:00"))
-    except ValueError as e:
-        raise ValueError(
-            f"Invalid timestamp format '{timestamp}'. "
-            "Expected ISO format like '2024-01-15T10:30:00' or '2024-01-15T10:30:00Z'"
-        ) from e
-
-
-def build_content_dict(
-    content: str,
-    context: str,
-    timestamp: str | None = None,
-) -> tuple[dict[str, Any], str | None]:
-    """Build a content dict for retain operations.
-
-    Args:
-        content: The memory content
-        context: Category for the memory
-        timestamp: Optional ISO timestamp
-
-    Returns:
-        Tuple of (content_dict, error_message). error_message is None if successful.
-    """
-    content_dict: dict[str, Any] = {"content": content, "context": context}
-
-    if timestamp:
-        try:
-            parsed_timestamp = parse_timestamp(timestamp)
-            content_dict["event_date"] = parsed_timestamp
-        except ValueError as e:
-            return {}, str(e)
-
-    return content_dict, None
-
-
-def register_mcp_tools(
-    mcp: FastMCP,
-    memory: MemoryEngine,
-    config: MCPToolsConfig,
-) -> None:
-    """Register MCP tools on a FastMCP server.
-
-    Args:
-        mcp: FastMCP server instance
-        memory: MemoryEngine instance
-        config: Tool configuration
-    """
-    tools_to_register = config.tools or {"retain", "recall", "reflect", "list_banks", "create_bank"}
-
-    if "retain" in tools_to_register:
-        _register_retain(mcp, memory, config)
-
-    if "recall" in tools_to_register:
-        _register_recall(mcp, memory, config)
-
-    if "reflect" in tools_to_register:
-        _register_reflect(mcp, memory, config)
-
-    if "list_banks" in tools_to_register:
-        _register_list_banks(mcp, memory, config)
-
-    if "create_bank" in tools_to_register:
-        _register_create_bank(mcp, memory, config)
-
-
-def _register_retain(mcp: FastMCP, memory: MemoryEngine, config: MCPToolsConfig) -> None:
-    """Register the retain tool."""
-    description = config.retain_description or DEFAULT_MCP_RETAIN_DESCRIPTION
-
-    if config.include_bank_id_param:
-        if config.retain_fire_and_forget:
-
-            @mcp.tool(description=description)
-            async def retain(
-                content: str,
-                context: str = "general",
-                timestamp: str | None = None,
-                bank_id: str | None = None,
-            ) -> dict:
-                """
-                Args:
-                    content: The fact/memory to store (be specific and include relevant details)
-                    context: Category for the memory (e.g., 'preferences', 'work', 'hobbies', 'family'). Default: 'general'
-                    timestamp: When this event/fact occurred (ISO format, e.g., '2024-01-15T10:30:00Z'). Useful for timeline tracking.
-                    bank_id: Optional bank to store in (defaults to session bank). Use for cross-bank operations.
-                """
-                import asyncio
-
-                target_bank = bank_id or config.bank_id_resolver()
-                if target_bank is None:
-                    return {"status": "error", "message": "No bank_id configured"}
-
-                content_dict, error = build_content_dict(content, context, timestamp)
-                if error:
-                    return {"status": "error", "message": error}
-
-                async def _retain():
-                    try:
-                        await memory.retain_batch_async(
-                            bank_id=target_bank,
-                            contents=[content_dict],
-                            request_context=RequestContext(),
-                        )
-                    except Exception as e:
-                        logger.error(f"Error storing memory: {e}", exc_info=True)
-
-                asyncio.create_task(_retain())
-                return {"status": "accepted", "message": "Memory storage initiated"}
-
-        else:
-
-            @mcp.tool(description=description)
-            async def retain(
-                content: str,
-                context: str = "general",
-                timestamp: str | None = None,
-                async_processing: bool = True,
-                bank_id: str | None = None,
-            ) -> str:
-                """
-                Args:
-                    content: The fact/memory to store (be specific and include relevant details)
-                    context: Category for the memory (e.g., 'preferences', 'work', 'hobbies', 'family'). Default: 'general'
-                    timestamp: When this event/fact occurred (ISO format, e.g., '2024-01-15T10:30:00Z'). Useful for timeline tracking.
-                    async_processing: If True, queue for background processing and return immediately. If False, wait for completion. Default: True
-                    bank_id: Optional bank to store in (defaults to session bank). Use for cross-bank operations.
-                """
-                try:
-                    target_bank = bank_id or config.bank_id_resolver()
-                    if target_bank is None:
-                        return "Error: No bank_id configured"
-
-                    content_dict, error = build_content_dict(content, context, timestamp)
-                    if error:
-                        return f"Error: {error}"
-
-                    contents = [content_dict]
-                    if async_processing:
-                        result = await memory.submit_async_retain(
-                            bank_id=target_bank, contents=contents, request_context=RequestContext()
-                        )
-                        return f"Memory queued for background processing (operation_id: {result.get('operation_id', 'N/A')})"
-                    else:
-                        await memory.retain_batch_async(
-                            bank_id=target_bank,
-                            contents=contents,
-                            request_context=RequestContext(),
-                        )
-                        return f"Memory stored successfully in bank '{target_bank}'"
-                except Exception as e:
-                    logger.error(f"Error storing memory: {e}", exc_info=True)
-                    return f"Error: {str(e)}"
-
-    else:
-        # No bank_id param - use fixed bank from resolver
-
-        @mcp.tool(description=description)
-        async def retain(
-            content: str,
-            context: str = "general",
-            timestamp: str | None = None,
-        ) -> dict:
-            """
-            Args:
-                content: The fact/memory to store (be specific and include relevant details)
-                context: Category for the memory (e.g., 'preferences', 'work', 'hobbies', 'family'). Default: 'general'
-                timestamp: When this event/fact occurred (ISO format, e.g., '2024-01-15T10:30:00Z'). Useful for timeline tracking.
-            """
-            import asyncio
-
-            target_bank = config.bank_id_resolver()
-            if target_bank is None:
-                return {"status": "error", "message": "No bank_id configured"}
-
-            content_dict, error = build_content_dict(content, context, timestamp)
-            if error:
-                return {"status": "error", "message": error}
-
-            async def _retain():
-                try:
-                    await memory.retain_batch_async(
-                        bank_id=target_bank,
-                        contents=[content_dict],
-                        request_context=RequestContext(),
-                    )
-                except Exception as e:
-                    logger.error(f"Error storing memory: {e}", exc_info=True)
-
-            asyncio.create_task(_retain())
-            return {"status": "accepted", "message": "Memory storage initiated"}
-
-
-def _register_recall(mcp: FastMCP, memory: MemoryEngine, config: MCPToolsConfig) -> None:
-    """Register the recall tool."""
-    description = config.recall_description or DEFAULT_MCP_RECALL_DESCRIPTION
-
-    if config.include_bank_id_param:
-
-        @mcp.tool(description=description)
-        async def recall(
-            query: str,
-            max_tokens: int = 4096,
-            bank_id: str | None = None,
-        ) -> str | dict:
-            """
-            Args:
-                query: Natural language search query (e.g., "user's food preferences", "what projects is user working on")
-                max_tokens: Maximum tokens to return in results (default: 4096)
-                bank_id: Optional bank to search in (defaults to session bank). Use for cross-bank operations.
-            """
-            try:
-                target_bank = bank_id or config.bank_id_resolver()
-                if target_bank is None:
-                    return "Error: No bank_id configured"
-
-                recall_result = await memory.recall_async(
-                    bank_id=target_bank,
-                    query=query,
-                    fact_type=list(VALID_RECALL_FACT_TYPES),
-                    budget=Budget.HIGH,
-                    max_tokens=max_tokens,
-                    request_context=RequestContext(),
-                )
-
-                return recall_result.model_dump_json(indent=2)
-            except Exception as e:
-                logger.error(f"Error searching: {e}", exc_info=True)
-                return f'{{"error": "{e}", "results": []}}'
-
-    else:
-
-        @mcp.tool(description=description)
-        async def recall(
-            query: str,
-            max_tokens: int = 4096,
-        ) -> dict:
-            """
-            Args:
-                query: Natural language search query (e.g., "user's food preferences", "what projects is user working on")
-                max_tokens: Maximum tokens to return in results (default: 4096)
-            """
-            try:
-                target_bank = config.bank_id_resolver()
-                if target_bank is None:
-                    return {"error": "No bank_id configured", "results": []}
-
-                recall_result = await memory.recall_async(
-                    bank_id=target_bank,
-                    query=query,
-                    fact_type=list(VALID_RECALL_FACT_TYPES),
-                    budget=Budget.HIGH,
-                    max_tokens=max_tokens,
-                    request_context=RequestContext(),
-                )
-
-                return recall_result.model_dump()
-            except Exception as e:
-                logger.error(f"Error searching: {e}", exc_info=True)
-                return {"error": str(e), "results": []}
-
-
-def _register_reflect(mcp: FastMCP, memory: MemoryEngine, config: MCPToolsConfig) -> None:
-    """Register the reflect tool."""
-
-    if config.include_bank_id_param:
-
-        @mcp.tool()
-        async def reflect(
-            query: str,
-            context: str | None = None,
-            budget: str = "low",
-            bank_id: str | None = None,
-        ) -> str:
-            """
-            Generate thoughtful analysis by synthesizing stored memories with the bank's personality.
-
-            WHEN TO USE THIS TOOL:
-            Use reflect when you need reasoned analysis, not just fact retrieval. This tool
-            thinks through the question using everything the bank knows and its personality traits.
-
-            EXAMPLES OF GOOD QUERIES:
-            - "What patterns have emerged in how I approach debugging?"
-            - "Based on my past decisions, what architectural style do I prefer?"
-            - "What might be the best approach for this problem given what you know about me?"
-            - "How should I prioritize these tasks based on my goals?"
-
-            HOW IT DIFFERS FROM RECALL:
-            - recall: Returns raw facts matching your search (fast lookup)
-            - reflect: Reasons across memories to form a synthesized answer (deeper analysis)
-
-            Use recall for "what did I say about X?" and reflect for "what should I do about X?"
-
-            Args:
-                query: The question or topic to reflect on
-                context: Optional context about why this reflection is needed
-                budget: Search budget - 'low', 'mid', or 'high' (default: 'low')
-                bank_id: Optional bank to reflect in (defaults to session bank). Use for cross-bank operations.
-            """
-            try:
-                target_bank = bank_id or config.bank_id_resolver()
-                if target_bank is None:
-                    return "Error: No bank_id configured"
-
-                budget_map = {"low": Budget.LOW, "mid": Budget.MID, "high": Budget.HIGH}
-                budget_enum = budget_map.get(budget.lower(), Budget.LOW)
-
-                reflect_result = await memory.reflect_async(
-                    bank_id=target_bank,
-                    query=query,
-                    budget=budget_enum,
-                    context=context,
-                    request_context=RequestContext(),
-                )
-
-                return reflect_result.model_dump_json(indent=2)
-            except Exception as e:
-                logger.error(f"Error reflecting: {e}", exc_info=True)
-                return f'{{"error": "{e}", "text": ""}}'
-
-    else:
-
-        @mcp.tool()
-        async def reflect(
-            query: str,
-            context: str | None = None,
-            budget: str = "low",
-        ) -> dict:
-            """
-            Generate thoughtful analysis by synthesizing stored memories with the bank's personality.
-
-            WHEN TO USE THIS TOOL:
-            Use reflect when you need reasoned analysis, not just fact retrieval. This tool
-            thinks through the question using everything the bank knows and its personality traits.
-
-            EXAMPLES OF GOOD QUERIES:
-            - "What patterns have emerged in how I approach debugging?"
-            - "Based on my past decisions, what architectural style do I prefer?"
-            - "What might be the best approach for this problem given what you know about me?"
-            - "How should I prioritize these tasks based on my goals?"
-
-            HOW IT DIFFERS FROM RECALL:
-            - recall: Returns raw facts matching your search (fast lookup)
-            - reflect: Reasons across memories to form a synthesized answer (deeper analysis)
-
-            Use recall for "what did I say about X?" and reflect for "what should I do about X?"
-
-            Args:
-                query: The question or topic to reflect on
-                context: Optional context about why this reflection is needed
-                budget: Search budget - 'low', 'mid', or 'high' (default: 'low')
-            """
-            try:
-                target_bank = config.bank_id_resolver()
-                if target_bank is None:
-                    return {"error": "No bank_id configured", "text": ""}
-
-                budget_map = {"low": Budget.LOW, "mid": Budget.MID, "high": Budget.HIGH}
-                budget_enum = budget_map.get(budget.lower(), Budget.LOW)
-
-                reflect_result = await memory.reflect_async(
-                    bank_id=target_bank,
-                    query=query,
-                    budget=budget_enum,
-                    context=context,
-                    request_context=RequestContext(),
-                )
-
-                return reflect_result.model_dump()
-            except Exception as e:
-                logger.error(f"Error reflecting: {e}", exc_info=True)
-                return {"error": str(e), "text": ""}
-
-
-def _register_list_banks(mcp: FastMCP, memory: MemoryEngine, config: MCPToolsConfig) -> None:
-    """Register the list_banks tool."""
-
-    @mcp.tool()
-    async def list_banks() -> str:
-        """
-        List all available memory banks.
-
-        Use this tool to discover what memory banks exist in the system.
-        Each bank is an isolated memory store (like a separate "brain").
-
-        Returns:
-            JSON list of banks with their IDs, names, dispositions, and missions.
-        """
-        try:
-            banks = await memory.list_banks(request_context=RequestContext())
-            return json.dumps({"banks": banks}, indent=2)
-        except Exception as e:
-            logger.error(f"Error listing banks: {e}", exc_info=True)
-            return f'{{"error": "{e}", "banks": []}}'
-
-
-def _register_create_bank(mcp: FastMCP, memory: MemoryEngine, config: MCPToolsConfig) -> None:
-    """Register the create_bank tool."""
-
-    @mcp.tool()
-    async def create_bank(bank_id: str, name: str | None = None, mission: str | None = None) -> str:
-        """
-        Create a new memory bank or get an existing one.
-
-        Memory banks are isolated stores - each one is like a separate "brain" for a user/agent.
-        Banks are auto-created with default settings if they don't exist.
-
-        Args:
-            bank_id: Unique identifier for the bank (e.g., 'user-123', 'agent-alpha')
-            name: Optional human-friendly name for the bank
-            mission: Optional mission describing who the agent is and what they're trying to accomplish
-        """
-        try:
-            # get_bank_profile auto-creates bank if it doesn't exist
-            profile = await memory.get_bank_profile(bank_id, request_context=RequestContext())
-
-            # Update name/mission if provided
-            if name is not None or mission is not None:
-                await memory.update_bank(
-                    bank_id,
-                    name=name,
-                    mission=mission,
-                    request_context=RequestContext(),
-                )
-                # Fetch updated profile
-                profile = await memory.get_bank_profile(bank_id, request_context=RequestContext())
-
-            # Serialize disposition if it's a Pydantic model
-            if "disposition" in profile and hasattr(profile["disposition"], "model_dump"):
-                profile["disposition"] = profile["disposition"].model_dump()
-            return json.dumps(profile, indent=2)
-        except Exception as e:
-            logger.error(f"Error creating bank: {e}", exc_info=True)
-            return f'{{"error": "{e}"}}'
@@ -28,15 +28,6 @@ from opentelemetry.sdk.resources import Resource
 if TYPE_CHECKING:
    import asyncpg

-
-def _get_tenant() -> str:
-    """Get current tenant (schema) from context for metrics labeling."""
-    # Import here to avoid circular imports
-    from hindsight_api.engine.memory_engine import get_current_schema
-
-    return get_current_schema()
-
-
 # Custom bucket boundaries for operation duration (in seconds)
 # Fine granularity in 0-30s range where most operations complete
 DURATION_BUCKETS = (0.1, 0.25, 0.5, 0.75, 1.0, 2.0, 3.0, 5.0, 7.5, 10.0, 15.0, 20.0, 30.0, 60.0, 120.0)
@@ -332,7 +323,6 @@ class MetricsCollector(MetricsCollectorBase):
            "operation": operation,
            "bank_id": bank_id,
            "source": source,
-            "tenant": _get_tenant(),
        }
        if budget:
            attributes["budget"] = budget
@@ -383,7 +373,6 @@ class MetricsCollector(MetricsCollectorBase):
            "model": model,
            "scope": scope,
            "success": str(success).lower(),
-            "tenant": _get_tenant(),
        }

        # Record duration
@@ -436,14 +425,10 @@ class MetricsCollector(MetricsCollectorBase):
            status_code = status_code_getter()
            status_class = f"{status_code // 100}xx"

-            # Get tenant from context (may be set during request processing)
-            tenant = _get_tenant()
-
            attributes = {
                **base_attributes,
                "status_code": str(status_code),
                "status_class": status_class,
-                "tenant": tenant,
            }

            # Record duration and count
@@ -95,6 +95,7 @@ class MemoryUnit(Base):
    mentioned_at: Mapped[datetime | None] = mapped_column(TIMESTAMP(timezone=True))  # When fact was mentioned
    fact_type: Mapped[str] = mapped_column(Text, nullable=False, server_default="world")
    confidence_score: Mapped[float | None] = mapped_column(Float)
+    access_count: Mapped[int] = mapped_column(Integer, server_default="0")
    unit_metadata: Mapped[dict] = mapped_column(
        "metadata", JSONB, server_default=sql_text("'{}'::jsonb")
    )  # User-defined metadata (str->str)
@@ -130,6 +131,7 @@ class MemoryUnit(Base):
        Index("idx_memory_units_document_id", "document_id"),
        Index("idx_memory_units_event_date", "event_date", postgresql_ops={"event_date": "DESC"}),
        Index("idx_memory_units_bank_date", "bank_id", "event_date", postgresql_ops={"event_date": "DESC"}),
+        Index("idx_memory_units_access_count", "access_count", postgresql_ops={"access_count": "DESC"}),
        Index("idx_memory_units_fact_type", "fact_type"),
        Index("idx_memory_units_bank_fact_type", "bank_id", "fact_type"),
        Index(
@@ -7,7 +7,6 @@ This module provides the ASGI app for uvicorn import string usage:
 For CLI usage, use the hindsight-api command instead.
 """

-import logging
 import os
 import warnings

@@ -18,12 +17,6 @@ warnings.filterwarnings("ignore", message="websockets.server.WebSocketServerProt
 from hindsight_api import MemoryEngine
 from hindsight_api.api import create_app
 from hindsight_api.config import get_config
-from hindsight_api.extensions import (
-    DefaultExtensionContext,
-    OperationValidatorExtension,
-    TenantExtension,
-    load_extension,
-)

 # Disable tokenizers parallelism to avoid warnings
 os.environ["TOKENIZERS_PARALLELISM"] = "false"
@@ -32,33 +25,10 @@ os.environ["TOKENIZERS_PARALLELISM"] = "false"
 config = get_config()
 config.configure_logging()

-# Load operation validator extension if configured
-operation_validator = load_extension("OPERATION_VALIDATOR", OperationValidatorExtension)
-if operation_validator:
-    logging.info(f"Loaded operation validator: {operation_validator.__class__.__name__}")
-
-# Load tenant extension if configured
-tenant_extension = load_extension("TENANT", TenantExtension)
-if tenant_extension:
-    logging.info(f"Loaded tenant extension: {tenant_extension.__class__.__name__}")
-
 # Create app at module level (required for uvicorn import string)
 # MemoryEngine reads configuration from environment variables automatically
 # Note: run_migrations=True by default, but migrations are idempotent so safe with workers
-_memory = MemoryEngine(
-    operation_validator=operation_validator,
-    tenant_extension=tenant_extension,
-    run_migrations=config.run_migrations_on_startup,
-)
-
-# Set extension context on tenant extension (needed for schema provisioning)
-if tenant_extension:
-    extension_context = DefaultExtensionContext(
-        database_url=config.database_url,
-        memory_engine=_memory,
-    )
-    tenant_extension.set_context(extension_context)
-    logging.info("Extension context set on tenant extension")
+_memory = MemoryEngine(run_migrations=config.run_migrations_on_startup)

 # Create unified app with both HTTP and optionally MCP
 app = create_app(
@@ -1,11 +0,0 @@
-"""
-Worker package for distributed task processing.
-
-This package provides:
- WorkerPoller: Polls PostgreSQL for pending tasks and executes them
- main: CLI entry point for hindsight-worker
-"""
-
-from .poller import WorkerPoller
-
-__all__ = ["WorkerPoller"]
@@ -1,296 +0,0 @@
-"""
-Command-line interface for Hindsight Worker.
-
-Run the worker with:
-    hindsight-worker
-
-Stop with Ctrl+C (graceful shutdown).
-"""
-
-import argparse
-import asyncio
-import atexit
-import logging
-import os
-import signal
-import socket
-import sys
-import warnings
-
-from ..config import get_config
-from ..engine.task_backend import SyncTaskBackend
-from .poller import WorkerPoller
-
-# Filter deprecation warnings from third-party libraries
-warnings.filterwarnings("ignore", message="websockets.legacy is deprecated")
-warnings.filterwarnings("ignore", message="websockets.server.WebSocketServerProtocol is deprecated")
-
-# Disable tokenizers parallelism to avoid warnings
-os.environ["TOKENIZERS_PARALLELISM"] = "false"
-
-logger = logging.getLogger(__name__)
-
-
-def create_worker_app(poller: WorkerPoller, memory):
-    """Create a minimal FastAPI app for worker metrics and health."""
-    from fastapi import FastAPI
-    from fastapi.responses import JSONResponse, Response
-    from prometheus_client import CONTENT_TYPE_LATEST, generate_latest
-
-    from ..metrics import create_metrics_collector, get_metrics_collector, initialize_metrics
-
-    app = FastAPI(
-        title="Hindsight Worker",
-        description="Worker process for distributed task execution",
-    )
-
-    # Initialize OpenTelemetry metrics
-    try:
-        prometheus_reader = initialize_metrics(service_name="hindsight-worker", service_version="1.0.0")
-        create_metrics_collector()
-        app.state.prometheus_reader = prometheus_reader
-        logger.info("Metrics initialized - available at /metrics endpoint")
-    except Exception as e:
-        logger.warning(f"Failed to initialize metrics: {e}. Metrics will be disabled.")
-        app.state.prometheus_reader = None
-
-    # Set up DB pool metrics if available
-    metrics_collector = get_metrics_collector()
-    if memory._pool is not None and hasattr(metrics_collector, "set_db_pool"):
-        metrics_collector.set_db_pool(memory._pool)
-        logger.info("DB pool metrics configured")
-
-    @app.get(
-        "/health",
-        summary="Health check endpoint",
-        description="Returns worker health status including database connectivity",
-        tags=["Monitoring"],
-    )
-    async def health_endpoint():
-        """Health check endpoint."""
-        health = await memory.health_check()
-        health["worker_id"] = poller.worker_id
-        health["is_shutdown"] = poller.is_shutdown
-        status_code = 200 if health.get("status") == "healthy" else 503
-        return JSONResponse(content=health, status_code=status_code)
-
-    @app.get(
-        "/metrics",
-        summary="Prometheus metrics endpoint",
-        description="Exports metrics in Prometheus format for scraping",
-        tags=["Monitoring"],
-    )
-    async def metrics_endpoint():
-        """Return Prometheus metrics."""
-        metrics_data = generate_latest()
-        return Response(content=metrics_data, media_type=CONTENT_TYPE_LATEST)
-
-    @app.get(
-        "/",
-        summary="Worker info",
-        description="Basic worker information",
-        tags=["Info"],
-    )
-    async def root():
-        """Return basic worker info."""
-        return {
-            "service": "hindsight-worker",
-            "worker_id": poller.worker_id,
-            "is_shutdown": poller.is_shutdown,
-        }
-
-    return app
-
-
-def main():
-    """Main entry point for the hindsight-worker CLI."""
-    # Load configuration from environment
-    config = get_config()
-
-    parser = argparse.ArgumentParser(
-        prog="hindsight-worker",
-        description="Hindsight Worker - distributed task processor",
-    )
-
-    # Worker options
-    parser.add_argument(
-        "--worker-id",
-        default=config.worker_id or socket.gethostname(),
-        help="Worker identifier (default: hostname, env: HINDSIGHT_API_WORKER_ID)",
-    )
-    parser.add_argument(
-        "--poll-interval",
-        type=int,
-        default=config.worker_poll_interval_ms,
-        help=f"Poll interval in milliseconds (default: {config.worker_poll_interval_ms}, env: HINDSIGHT_API_WORKER_POLL_INTERVAL_MS)",
-    )
-    parser.add_argument(
-        "--batch-size",
-        type=int,
-        default=config.worker_batch_size,
-        help=f"Tasks to claim per poll (default: {config.worker_batch_size}, env: HINDSIGHT_API_WORKER_BATCH_SIZE)",
-    )
-    parser.add_argument(
-        "--max-retries",
-        type=int,
-        default=config.worker_max_retries,
-        help=f"Max retries before marking failed (default: {config.worker_max_retries}, env: HINDSIGHT_API_WORKER_MAX_RETRIES)",
-    )
-
-    # HTTP server options
-    parser.add_argument(
-        "--http-port",
-        type=int,
-        default=config.worker_http_port,
-        help=f"HTTP port for metrics/health endpoints (default: {config.worker_http_port}, env: HINDSIGHT_API_WORKER_HTTP_PORT)",
-    )
-    parser.add_argument(
-        "--http-host",
-        default="0.0.0.0",
-        help="HTTP host to bind (default: 0.0.0.0)",
-    )
-
-    # Logging options
-    parser.add_argument(
-        "--log-level",
-        default=config.log_level,
-        choices=["critical", "error", "warning", "info", "debug", "trace"],
-        help=f"Log level (default: {config.log_level}, env: HINDSIGHT_API_LOG_LEVEL)",
-    )
-
-    args = parser.parse_args()
-
-    # Configure logging
-    config.configure_logging()
-
-    # Import MemoryEngine here to avoid circular imports
-    from .. import MemoryEngine
-
-    print(f"Starting Hindsight Worker: {args.worker_id}")
-    print(f"  Poll interval: {args.poll_interval}ms")
-    print(f"  Batch size: {args.batch_size}")
-    print(f"  Max retries: {args.max_retries}")
-    print(f"  HTTP server: {args.http_host}:{args.http_port}")
-    print()
-
-    # Global references for cleanup
-    memory = None
-    poller = None
-
-    async def run():
-        nonlocal memory, poller
-        import uvicorn
-
-        from ..extensions import TenantExtension, load_extension
-
-        # Initialize MemoryEngine
-        # Workers use SyncTaskBackend because they execute tasks directly,
-        # they don't need to store tasks (they poll from DB)
-        memory = MemoryEngine(
-            run_migrations=False,  # Workers don't run migrations
-            task_backend=SyncTaskBackend(),
-        )
-
-        await memory.initialize()
-
-        print(f"Database connected: {config.database_url}")
-
-        # Load tenant extension for dynamic schema discovery
-        tenant_extension = load_extension("TENANT", TenantExtension)
-
-        if tenant_extension:
-            print("Tenant extension loaded - schemas will be discovered dynamically on each poll")
-        else:
-            print("No tenant extension configured, using public schema only")
-
-        # Create a single poller that handles all schemas dynamically
-        poller = WorkerPoller(
-            pool=memory._pool,
-            worker_id=args.worker_id,
-            executor=memory.execute_task,
-            poll_interval_ms=args.poll_interval,
-            batch_size=args.batch_size,
-            max_retries=args.max_retries,
-            tenant_extension=tenant_extension,
-        )
-
-        # Create the HTTP app for metrics/health
-        app = create_worker_app(poller, memory)
-
-        # Setup signal handlers for graceful shutdown
-        shutdown_requested = asyncio.Event()
-
-        def signal_handler(signum, frame):
-            print(f"\nReceived signal {signum}, initiating graceful shutdown...")
-            shutdown_requested.set()
-
-        signal.signal(signal.SIGINT, signal_handler)
-        signal.signal(signal.SIGTERM, signal_handler)
-
-        # Create uvicorn config and server
-        uvicorn_config = uvicorn.Config(
-            app,
-            host=args.http_host,
-            port=args.http_port,
-            log_level="info",  # Reduce uvicorn noise
-            access_log=False,
-        )
-        server = uvicorn.Server(uvicorn_config)
-
-        # Run the poller and HTTP server concurrently
-        poller_task = asyncio.create_task(poller.run())
-        http_task = asyncio.create_task(server.serve())
-
-        print(f"Worker started. Metrics available at http://{args.http_host}:{args.http_port}/metrics")
-
-        # Wait for shutdown signal
-        await shutdown_requested.wait()
-
-        # Graceful shutdown
-        print("Shutting down HTTP server...")
-        server.should_exit = True
-
-        print("Waiting for poller to finish...")
-        await poller.shutdown_graceful(timeout=30.0)
-        poller_task.cancel()
-        try:
-            await poller_task
-        except asyncio.CancelledError:
-            pass
-
-        # Wait for HTTP server to finish
-        try:
-            await asyncio.wait_for(http_task, timeout=5.0)
-        except asyncio.TimeoutError:
-            http_task.cancel()
-            try:
-                await http_task
-            except asyncio.CancelledError:
-                pass
-
-        # Close memory engine
-        await memory.close()
-        print("Worker shutdown complete")
-
-    def cleanup():
-        """Synchronous cleanup for atexit."""
-        if memory is not None and memory._pg0 is not None:
-            try:
-                loop = asyncio.new_event_loop()
-                loop.run_until_complete(memory._pg0.stop())
-                loop.close()
-                print("\npg0 stopped.")
-            except Exception as e:
-                print(f"\nError stopping pg0: {e}")
-
-    atexit.register(cleanup)
-
-    try:
-        asyncio.run(run())
-    except KeyboardInterrupt:
-        print("\nWorker interrupted")
-        sys.exit(0)
-
-
-if __name__ == "__main__":
-    main()
@@ -1,486 +0,0 @@
-"""
-Worker poller for distributed task execution.
-
-Polls PostgreSQL for pending tasks and executes them using
-FOR UPDATE SKIP LOCKED for safe concurrent claiming.
-"""
-
-import asyncio
-import json
-import logging
-import time
-import traceback
-from collections.abc import Awaitable, Callable
-from dataclasses import dataclass
-from typing import TYPE_CHECKING, Any
-
-if TYPE_CHECKING:
-    import asyncpg
-
-    from hindsight_api.extensions.tenant import TenantExtension
-
-logger = logging.getLogger(__name__)
-
-# Progress logging interval in seconds
-PROGRESS_LOG_INTERVAL = 30
-
-
-def fq_table(table: str, schema: str | None = None) -> str:
-    """Get fully-qualified table name with optional schema prefix."""
-    if schema:
-        return f'"{schema}".{table}'
-    return table
-
-
-@dataclass
-class ClaimedTask:
-    """A task claimed from the database with its schema context."""
-
-    operation_id: str
-    task_dict: dict[str, Any]
-    schema: str | None
-
-
-class WorkerPoller:
-    """
-    Polls PostgreSQL for pending tasks and executes them.
-
-    Uses FOR UPDATE SKIP LOCKED for safe distributed claiming,
-    allowing multiple workers to process tasks without conflicts.
-
-    Supports dynamic multi-tenant discovery via tenant_extension.
-    """
-
-    def __init__(
-        self,
-        pool: "asyncpg.Pool",
-        worker_id: str,
-        executor: Callable[[dict[str, Any]], Awaitable[None]],
-        poll_interval_ms: int = 500,
-        batch_size: int = 10,
-        max_retries: int = 3,
-        schema: str | None = None,
-        tenant_extension: "TenantExtension | None" = None,
-    ):
-        """
-        Initialize the worker poller.
-
-        Args:
-            pool: asyncpg connection pool
-            worker_id: Unique identifier for this worker
-            executor: Async function to execute tasks (typically MemoryEngine.execute_task)
-            poll_interval_ms: Interval between polls when no tasks found (milliseconds)
-            batch_size: Maximum number of tasks to claim per poll cycle
-            max_retries: Maximum retry attempts before marking task as failed
-            schema: Database schema for single-tenant support (ignored if tenant_extension is set)
-            tenant_extension: Extension for dynamic multi-tenant discovery. If set, list_tenants()
-                            is called on each poll cycle to discover schemas dynamically.
-        """
-        self._pool = pool
-        self._worker_id = worker_id
-        self._executor = executor
-        self._poll_interval_ms = poll_interval_ms
-        self._batch_size = batch_size
-        self._max_retries = max_retries
-        self._schema = schema
-        self._tenant_extension = tenant_extension
-        self._shutdown = asyncio.Event()
-        self._current_tasks: set[asyncio.Task] = set()
-        self._in_flight_count = 0
-        self._in_flight_lock = asyncio.Lock()
-        self._last_progress_log = 0.0
-        self._tasks_completed_since_log = 0
-        # Track active tasks locally: operation_id -> (op_type, bank_id, schema)
-        self._active_tasks: dict[str, tuple[str, str, str | None]] = {}
-
-    async def _get_schemas(self) -> list[str | None]:
-        """Get list of schemas to poll. Returns [None] for public schema."""
-        if self._tenant_extension is not None:
-            tenants = await self._tenant_extension.list_tenants()
-            # Convert "public" to None for SQL compatibility, keep others as-is
-            return [t.schema if t.schema != "public" else None for t in tenants]
-        # Single schema mode
-        return [self._schema]
-
-    async def claim_batch(self) -> list[ClaimedTask]:
-        """
-        Claim up to batch_size pending tasks atomically across all tenant schemas.
-
-        Uses FOR UPDATE SKIP LOCKED to ensure no conflicts with other workers.
-
-        For consolidation tasks specifically, skips pending tasks if there's already
-        a processing consolidation for the same bank (to avoid duplicate work).
-
-        If tenant_extension is configured, dynamically discovers schemas on each call.
-
-        Returns:
-            List of ClaimedTask objects containing operation_id, task_dict, and schema
-        """
-        schemas = await self._get_schemas()
-        all_tasks: list[ClaimedTask] = []
-        remaining_batch = self._batch_size
-
-        for schema in schemas:
-            if remaining_batch <= 0:
-                break
-
-            tasks = await self._claim_batch_for_schema(schema, remaining_batch)
-            all_tasks.extend(tasks)
-            remaining_batch -= len(tasks)
-
-        return all_tasks
-
-    async def _claim_batch_for_schema(self, schema: str | None, limit: int) -> list[ClaimedTask]:
-        """Claim tasks from a specific schema."""
-        table = fq_table("async_operations", schema)
-
-        async with self._pool.acquire() as conn:
-            async with conn.transaction():
-                # Select and lock pending tasks
-                # For consolidation: skip if same bank already has one processing
-                rows = await conn.fetch(
-                    f"""
-                    SELECT operation_id, task_payload
-                    FROM {table} AS pending
-                    WHERE status = 'pending' AND task_payload IS NOT NULL
-                    AND (
-                        -- Non-consolidation tasks: always claimable
-                        operation_type != 'consolidation'
-                        OR
-                        -- Consolidation: only if no other consolidation processing for same bank
-                        NOT EXISTS (
-                            SELECT 1 FROM {table} AS processing
-                            WHERE processing.bank_id = pending.bank_id
-                            AND processing.operation_type = 'consolidation'
-                            AND processing.status = 'processing'
-                        )
-                    )
-                    ORDER BY created_at
-                    LIMIT $1
-                    FOR UPDATE SKIP LOCKED
-                    """,
-                    limit,
-                )
-
-                if not rows:
-                    return []
-
-                # Claim the tasks by updating status and worker_id
-                operation_ids = [row["operation_id"] for row in rows]
-                await conn.execute(
-                    f"""
-                    UPDATE {table}
-                    SET status = 'processing', worker_id = $1, claimed_at = now(), updated_at = now()
-                    WHERE operation_id = ANY($2)
-                    """,
-                    self._worker_id,
-                    operation_ids,
-                )
-
-                # Parse and return task payloads with schema context
-                return [
-                    ClaimedTask(
-                        operation_id=str(row["operation_id"]),
-                        task_dict=json.loads(row["task_payload"]),
-                        schema=schema,
-                    )
-                    for row in rows
-                ]
-
-    async def _mark_completed(self, operation_id: str, schema: str | None):
-        """Mark a task as completed."""
-        table = fq_table("async_operations", schema)
-        await self._pool.execute(
-            f"""
-            UPDATE {table}
-            SET status = 'completed', completed_at = now(), updated_at = now()
-            WHERE operation_id = $1
-            """,
-            operation_id,
-        )
-
-    async def _mark_failed(self, operation_id: str, error_message: str, schema: str | None):
-        """Mark a task as failed with error message."""
-        table = fq_table("async_operations", schema)
-        # Truncate error message if too long (max 5000 chars in schema)
-        error_message = error_message[:5000] if len(error_message) > 5000 else error_message
-        await self._pool.execute(
-            f"""
-            UPDATE {table}
-            SET status = 'failed', error_message = $2, completed_at = now(), updated_at = now()
-            WHERE operation_id = $1
-            """,
-            operation_id,
-            error_message,
-        )
-
-    async def _retry_or_fail(self, operation_id: str, error_message: str, schema: str | None):
-        """Increment retry count or mark as failed if max retries exceeded."""
-        table = fq_table("async_operations", schema)
-
-        # Get current retry count
-        row = await self._pool.fetchrow(
-            f"SELECT retry_count FROM {table} WHERE operation_id = $1",
-            operation_id,
-        )
-
-        if row is None:
-            logger.warning(f"Operation {operation_id} not found, cannot retry")
-            return
-
-        retry_count = row["retry_count"]
-
-        if retry_count >= self._max_retries:
-            # Max retries exceeded, mark as failed
-            await self._mark_failed(
-                operation_id, f"Max retries ({self._max_retries}) exceeded. Last error: {error_message}", schema
-            )
-            logger.error(f"Task {operation_id} failed after {retry_count} retries")
-        else:
-            # Increment retry and reset to pending
-            await self._pool.execute(
-                f"""
-                UPDATE {table}
-                SET status = 'pending', worker_id = NULL, claimed_at = NULL,
-                    retry_count = retry_count + 1, updated_at = now()
-                WHERE operation_id = $1
-                """,
-                operation_id,
-            )
-            logger.warning(f"Task {operation_id} failed, will retry (attempt {retry_count + 1}/{self._max_retries})")
-
-    async def execute_task(self, task: ClaimedTask):
-        """Execute a single task and update its status."""
-        task_type = task.task_dict.get("type", "unknown")
-        bank_id = task.task_dict.get("bank_id", "unknown")
-
-        # Track this task as active
-        async with self._in_flight_lock:
-            self._active_tasks[task.operation_id] = (task_type, bank_id, task.schema)
-
-        try:
-            schema_info = f", schema={task.schema}" if task.schema else ""
-            logger.debug(f"Executing task {task.operation_id} (type={task_type}, bank={bank_id}{schema_info})")
-            # Pass schema to executor so it can set the correct context
-            if task.schema:
-                task.task_dict["_schema"] = task.schema
-            await self._executor(task.task_dict)
-            await self._mark_completed(task.operation_id, task.schema)
-            logger.debug(f"Task {task.operation_id} completed successfully")
-        except Exception as e:
-            error_msg = f"{type(e).__name__}: {e}\n{traceback.format_exc()}"
-            logger.error(f"Task {task.operation_id} failed: {e}")
-            await self._retry_or_fail(task.operation_id, error_msg, task.schema)
-        finally:
-            # Remove from active tasks
-            async with self._in_flight_lock:
-                self._active_tasks.pop(task.operation_id, None)
-
-    async def recover_own_tasks(self) -> int:
-        """
-        Recover tasks that were assigned to this worker but not completed.
-
-        This handles the case where a worker crashes while processing tasks.
-        On startup, we reset any tasks stuck in 'processing' for this worker_id
-        back to 'pending' so they can be picked up again.
-
-        If tenant_extension is configured, recovers across all tenant schemas.
-
-        Returns:
-            Number of tasks recovered
-        """
-        schemas = await self._get_schemas()
-        total_count = 0
-
-        for schema in schemas:
-            table = fq_table("async_operations", schema)
-
-            result = await self._pool.execute(
-                f"""
-                UPDATE {table}
-                SET status = 'pending', worker_id = NULL, claimed_at = NULL, updated_at = now()
-                WHERE status = 'processing' AND worker_id = $1
-                """,
-                self._worker_id,
-            )
-
-            # Parse "UPDATE N" to get count
-            count = int(result.split()[-1]) if result else 0
-            total_count += count
-
-        if total_count > 0:
-            logger.info(f"Worker {self._worker_id} recovered {total_count} stale tasks from previous run")
-        return total_count
-
-    async def run(self):
-        """
-        Main polling loop.
-
-        Continuously polls for pending tasks, claims them, and executes them
-        until shutdown is signaled.
-
-        If tenant_extension is configured, dynamically discovers schemas on each poll.
-        """
-        # Recover any tasks from a previous crash before starting
-        await self.recover_own_tasks()
-
-        logger.info(f"Worker {self._worker_id} starting polling loop")
-
-        while not self._shutdown.is_set():
-            try:
-                # Claim a batch of tasks (across all tenant schemas if configured)
-                tasks = await self.claim_batch()
-
-                if tasks:
-                    # Log batch info
-                    task_types: dict[str, int] = {}
-                    schemas_seen: set[str | None] = set()
-                    for task in tasks:
-                        t = task.task_dict.get("type", "unknown")
-                        task_types[t] = task_types.get(t, 0) + 1
-                        schemas_seen.add(task.schema)
-                    types_str = ", ".join(f"{k}:{v}" for k, v in task_types.items())
-                    schemas_str = ", ".join(s or "public" for s in schemas_seen)
-                    logger.info(
-                        f"Worker {self._worker_id} claimed {len(tasks)} tasks: {types_str} (schemas: {schemas_str})"
-                    )
-
-                    # Track in-flight tasks
-                    async with self._in_flight_lock:
-                        self._in_flight_count += len(tasks)
-
-                    # Execute tasks concurrently
-                    try:
-                        await asyncio.gather(
-                            *[self.execute_task(task) for task in tasks],
-                            return_exceptions=True,
-                        )
-                    finally:
-                        async with self._in_flight_lock:
-                            self._in_flight_count -= len(tasks)
-                else:
-                    # No tasks found, wait before polling again
-                    try:
-                        await asyncio.wait_for(
-                            self._shutdown.wait(),
-                            timeout=self._poll_interval_ms / 1000,
-                        )
-                    except asyncio.TimeoutError:
-                        pass  # Normal timeout, continue polling
-
-                # Log progress stats periodically
-                await self._log_progress_if_due()
-
-            except asyncio.CancelledError:
-                logger.info(f"Worker {self._worker_id} polling loop cancelled")
-                break
-            except Exception as e:
-                logger.error(f"Worker {self._worker_id} error in polling loop: {e}")
-                traceback.print_exc()
-                # Backoff on error
-                await asyncio.sleep(1)
-
-        logger.info(f"Worker {self._worker_id} polling loop stopped")
-
-    async def shutdown_graceful(self, timeout: float = 30.0):
-        """
-        Signal shutdown and wait for current tasks to complete.
-
-        Args:
-            timeout: Maximum time to wait for in-flight tasks (seconds)
-        """
-        logger.info(f"Worker {self._worker_id} initiating graceful shutdown")
-        self._shutdown.set()
-
-        # Wait for in-flight tasks to complete
-        start_time = asyncio.get_event_loop().time()
-        while asyncio.get_event_loop().time() - start_time < timeout:
-            async with self._in_flight_lock:
-                in_flight = self._in_flight_count
-
-            if in_flight == 0:
-                logger.info(f"Worker {self._worker_id} graceful shutdown complete")
-                return
-
-            logger.info(f"Worker {self._worker_id} waiting for {in_flight} in-flight tasks")
-            await asyncio.sleep(0.5)
-
-        logger.warning(f"Worker {self._worker_id} shutdown timeout after {timeout}s")
-
-    async def _log_progress_if_due(self):
-        """Log progress stats every PROGRESS_LOG_INTERVAL seconds."""
-        now = time.time()
-        if now - self._last_progress_log < PROGRESS_LOG_INTERVAL:
-            return
-
-        self._last_progress_log = now
-
-        try:
-            # Get local active tasks (this worker only)
-            async with self._in_flight_lock:
-                in_flight = self._in_flight_count
-                active_tasks = dict(self._active_tasks)  # Copy to avoid holding lock
-
-            # Build local processing breakdown grouped by (op_type, bank_id)
-            task_groups: dict[tuple[str, str], int] = {}
-            for op_type, bank_id, _ in active_tasks.values():
-                key = (op_type, bank_id)
-                task_groups[key] = task_groups.get(key, 0) + 1
-
-            processing_info = [f"{op}:{bank}({cnt})" for (op, bank), cnt in task_groups.items()]
-            processing_str = ", ".join(processing_info[:10]) if processing_info else "none"
-            if len(processing_info) > 10:
-                processing_str += f" +{len(processing_info) - 10} more"
-
-            # Get global stats from DB across all schemas
-            schemas = await self._get_schemas()
-            global_pending = 0
-            all_worker_counts: dict[str, int] = {}
-
-            async with self._pool.acquire() as conn:
-                for schema in schemas:
-                    table = fq_table("async_operations", schema)
-
-                    row = await conn.fetchrow(f"SELECT COUNT(*) as count FROM {table} WHERE status = 'pending'")
-                    global_pending += row["count"] if row else 0
-
-                    # Get processing breakdown by worker
-                    worker_rows = await conn.fetch(
-                        f"""
-                        SELECT worker_id, COUNT(*) as count
-                        FROM {table}
-                        WHERE status = 'processing'
-                        GROUP BY worker_id
-                        """
-                    )
-                    for wr in worker_rows:
-                        wid = wr["worker_id"] or "unknown"
-                        all_worker_counts[wid] = all_worker_counts.get(wid, 0) + wr["count"]
-
-            # Format other workers' processing counts
-            other_workers = []
-            for wid, cnt in all_worker_counts.items():
-                if wid != self._worker_id:
-                    other_workers.append(f"{wid}:{cnt}")
-            others_str = ", ".join(other_workers) if other_workers else "none"
-
-            schemas_str = ", ".join(s or "public" for s in schemas)
-            logger.info(
-                f"[WORKER_STATS] worker={self._worker_id} in_flight={in_flight} | "
-                f"global: pending={global_pending} (schemas: {schemas_str}) | "
-                f"others: {others_str} | "
-                f"my_active: {processing_str}"
-            )
-
-        except Exception as e:
-            logger.debug(f"Failed to log progress stats: {e}")
-
-    @property
-    def worker_id(self) -> str:
-        """Get the worker ID."""
-        return self._worker_id
-
-    @property
-    def is_shutdown(self) -> bool:
-        """Check if shutdown has been signaled."""
-        return self._shutdown.is_set()
@@ -4,7 +4,7 @@ build-backend = "hatchling.build"

 [project]
 name = "hindsight-api"
-version = "0.3.0"
+version = "0.2.1"
 description = "Hindsight: Agent Memory That Works Like Human Memory"
 readme = "README.md"
 requires-python = ">=3.11"
@@ -25,7 +25,7 @@ dependencies = [
    "psycopg2-binary>=2.9.11",
    "tiktoken>=0.12.0",
    "httpx>=0.27.0",
-    "fastmcp>=2.14.0",  # CVE-2025-66416
+    "fastmcp>=2.3.0",
    "pg0-embedded>=0.11.0",
    "python-dateutil>=2.8.0",
    "opentelemetry-api>=1.20.0",
@@ -39,17 +39,10 @@ dependencies = [
    "cohere>=5.0.0",
    "flashrank>=0.2.0",
    # Local ML models for embeddings/reranking - can be excluded in Docker with INCLUDE_LOCAL_MODELS=false
-    "sentence-transformers>=3.3.0",
-    "transformers>=4.53.0",  # Security fixes for ReDoS vulnerabilities
-    "torch>=2.6.0",  # CVE fix for remote code execution
+    "sentence-transformers>=3.0.0,<3.3.0",
+    "transformers>=4.30.0,<4.46.0",
+    "torch>=2.0.0",
    "uvloop>=0.22.1",
-    # Transitive dependency security fixes
-    "pyasn1>=0.6.2",  # DoS vulnerability fix
-    "urllib3>=2.6.3",  # Decompression-bomb safeguards bypass fix
-    "langchain-core>=1.2.5",  # Serialization injection vulnerability fix
-    "filelock>=3.20.1",  # TOCTOU race condition fix
-    "authlib>=1.6.6",  # Account takeover vulnerability fix
-    "aiohttp>=3.13.3",  # Multiple DoS vulnerabilities
 ]

 [project.optional-dependencies]
@@ -58,12 +51,11 @@ test = [
    "pytest-asyncio>=0.21.0",
    "pytest-timeout>=2.4.0",
    "pytest-xdist>=3.0.0",
-    "filelock>=3.20.1",  # TOCTOU race condition fix
+    "filelock>=3.0.0",
 ]

 [project.scripts]
 hindsight-api = "hindsight_api.main:main"
-hindsight-worker = "hindsight_api.worker.main:main"
 hindsight-local-mcp = "hindsight_api.mcp_local:main"
 hindsight-admin = "hindsight_api.admin.cli:main"

@@ -105,7 +97,7 @@ dev = [
    "pytest-timeout>=2.4.0",
    "pytest-xdist>=3.8.0",
    "python-dotenv>=1.2.1",
-    "filelock>=3.20.1",  # TOCTOU race condition fix
+    "filelock>=3.0.0",
    "ruff>=0.8.0",
    "ty>=0.0.1",
 ]
@@ -12,7 +12,6 @@ from hindsight_api import MemoryEngine, LLMConfig, LocalSTEmbeddings, RequestCon

 from hindsight_api.engine.cross_encoder import LocalSTCrossEncoder
 from hindsight_api.engine.query_analyzer import DateparserQueryAnalyzer
-from hindsight_api.engine.task_backend import SyncTaskBackend
 from hindsight_api.pg0 import EmbeddedPostgres

 # Default pg0 instance configuration for tests
@@ -116,65 +115,16 @@ def llm_config():


@pytest.fixture(scope="session")
-def embeddings(tmp_path_factory, worker_id):
-    """
-    Session-scoped embeddings fixture with filelock to prevent race conditions.
+def embeddings():

-    When pytest-xdist runs multiple workers in parallel, they all try to load
-    models from the HuggingFace cache simultaneously, which can cause race
-    conditions and meta tensor errors. We use a filelock to serialize model
-    initialization across workers.
-    """
-    # Get shared temp dir for coordination between xdist workers
-    if worker_id == "master":
-        root_tmp_dir = tmp_path_factory.getbasetemp()
-    else:
-        root_tmp_dir = tmp_path_factory.getbasetemp().parent
+    return LocalSTEmbeddings()

-    lock_file = root_tmp_dir / "embeddings_init.lock"
-
-    emb = LocalSTEmbeddings()
-
-    # Serialize model initialization across workers
-    with filelock.FileLock(str(lock_file)):
-        loop = asyncio.new_event_loop()
-        try:
-            loop.run_until_complete(emb.initialize())
-        finally:
-            loop.close()
-
-    return emb


@pytest.fixture(scope="session")
-def cross_encoder(tmp_path_factory, worker_id):
-    """
-    Session-scoped cross-encoder fixture with filelock to prevent race conditions.
+def cross_encoder():

-    When pytest-xdist runs multiple workers in parallel, they all try to load
-    models from the HuggingFace cache simultaneously, which can cause race
-    conditions and meta tensor errors. We use a filelock to serialize model
-    initialization across workers.
-    """
-    # Get shared temp dir for coordination between xdist workers
-    if worker_id == "master":
-        root_tmp_dir = tmp_path_factory.getbasetemp()
-    else:
-        root_tmp_dir = tmp_path_factory.getbasetemp().parent
-
-    lock_file = root_tmp_dir / "cross_encoder_init.lock"
-
-    ce = LocalSTCrossEncoder()
-
-    # Serialize model initialization across workers
-    with filelock.FileLock(str(lock_file)):
-        loop = asyncio.new_event_loop()
-        try:
-            loop.run_until_complete(ce.initialize())
-        finally:
-            loop.close()
-
-    return ce
+    return LocalSTCrossEncoder()

@pytest.fixture(scope="session")
 def query_analyzer():
@@ -197,7 +147,6 @@ async def memory(pg0_db_url, embeddings, cross_encoder, query_analyzer):
    Uses pg0_db_url (a postgresql:// URL) directly, so MemoryEngine won't try to
    manage pg0 lifecycle - that's handled by the session-scoped pg0_db_url fixture.
    Migrations are disabled here since they're run once at session scope in pg0_db_url.
-    Uses SyncTaskBackend so async tasks execute immediately (no worker needed).
    """
    mem = MemoryEngine(
        db_url=pg0_db_url,  # Direct postgresql:// URL, not pg0://
@@ -211,7 +160,6 @@ async def memory(pg0_db_url, embeddings, cross_encoder, query_analyzer):
        pool_min_size=1,
        pool_max_size=5,
        run_migrations=False,  # Migrations already run at session scope
-        task_backend=SyncTaskBackend(),  # Execute tasks immediately in tests
    )
    await mem.initialize()
    yield mem
@@ -1,5 +1,5 @@
 """
-Tests for agent management API (profile, disposition).
+Tests for agent management API (profile, disposition, background).
 """
 import pytest
 import uuid
@@ -25,12 +25,15 @@ class TestAgentProfile:

        assert profile is not None
        assert "disposition" in profile
+        assert "background" in profile

        disposition = profile["disposition"]
        assert disposition.skepticism == 3
        assert disposition.literalism == 3
        assert disposition.empathy == 3

+        assert profile["background"] == ""
+
    @pytest.mark.asyncio
    async def test_update_agent_disposition(self, memory: MemoryEngine, request_context):
        """Test updating agent disposition traits."""
@@ -73,10 +76,63 @@ class TestAgentProfile:
        for agent in agents:
            assert "bank_id" in agent
            assert "disposition" in agent
+            assert "background" in agent
            assert "created_at" in agent
            assert "updated_at" in agent


+class TestAgentBackground:
+    """Tests for agent background management."""
+
+    @pytest.mark.asyncio
+    async def test_merge_agent_background(self, memory: MemoryEngine, request_context):
+        """Test merging agent background information."""
+        bank_id = unique_agent_id("test_profile_merge")
+
+        profile = await memory.get_bank_profile(bank_id, request_context=request_context)
+        assert profile["background"] == ""
+
+        result1 = await memory.merge_bank_background(
+            bank_id,
+            "I was born in Texas",
+            update_disposition=False,
+            request_context=request_context,
+        )
+        assert "Texas" in result1["background"]
+
+        result2 = await memory.merge_bank_background(
+            bank_id,
+            "I have 10 years of startup experience",
+            update_disposition=False,
+            request_context=request_context,
+        )
+        assert "Texas" in result2["background"] or "startup" in result2["background"]
+
+        final_profile = await memory.get_bank_profile(bank_id, request_context=request_context)
+        assert final_profile["background"] != ""
+
+    @pytest.mark.asyncio
+    async def test_merge_background_handles_conflicts(self, memory: MemoryEngine, request_context):
+        """Test that merging background handles conflicts (new overwrites old)."""
+        bank_id = unique_agent_id("test_profile_conflict")
+
+        result1 = await memory.merge_bank_background(
+            bank_id,
+            "I was born in Colorado",
+            update_disposition=False,
+            request_context=request_context,
+        )
+        assert "Colorado" in result1["background"]
+
+        result2 = await memory.merge_bank_background(
+            bank_id,
+            "You were born in Texas",
+            update_disposition=False,
+            request_context=request_context,
+        )
+        assert "Texas" in result2["background"]
+
+
 class TestAgentEndpoint:
    """Tests for agent PUT endpoint logic."""

@@ -91,6 +147,7 @@ class TestAgentEndpoint:
                literalism=5,
                empathy=2
            ),
+            background="I am a creative software engineer"
        )

        profile = await memory.get_bank_profile(bank_id, request_context=request_context)
@@ -102,10 +159,55 @@ class TestAgentEndpoint:
                request_context=request_context,
            )

+        if request.background is not None:
+            pool = await memory._get_pool()
+            async with pool.acquire() as conn:
+                await conn.execute(
+                    """
+                    UPDATE banks
+                    SET background = $2,
+                        updated_at = NOW()
+                    WHERE bank_id = $1
+                    """,
+                    bank_id,
+                    request.background
+                )
+
        final_profile = await memory.get_bank_profile(bank_id, request_context=request_context)

        assert final_profile["disposition"].skepticism == 4
        assert final_profile["disposition"].literalism == 5
+        assert final_profile["background"] == "I am a creative software engineer"
+
+    @pytest.mark.asyncio
+    async def test_put_agent_partial_update(self, memory: MemoryEngine, request_context):
+        """Test updating only background."""
+        bank_id = unique_agent_id("test_put_partial")
+
+        request = CreateBankRequest(
+            background="I am a data scientist"
+        )
+
+        profile = await memory.get_bank_profile(bank_id, request_context=request_context)
+
+        if request.background is not None:
+            pool = await memory._get_pool()
+            async with pool.acquire() as conn:
+                await conn.execute(
+                    """
+                    UPDATE banks
+                    SET background = $2,
+                        updated_at = NOW()
+                    WHERE bank_id = $1
+                    """,
+                    bank_id,
+                    request.background
+                )
+
+        final_profile = await memory.get_bank_profile(bank_id, request_context=request_context)
+
+        assert final_profile["disposition"].skepticism == 3  # Default
+        assert final_profile["background"] == "I am a data scientist"


 class TestAgentDispositionIntegration:
@@ -123,6 +225,13 @@ class TestAgentDispositionIntegration:
        }
        await memory.update_bank_disposition(bank_id, disposition, request_context=request_context)

+        await memory.merge_bank_background(
+            bank_id,
+            "I am a creative artist who values innovation over tradition",
+            update_disposition=False,
+            request_context=request_context,
+        )
+
        await memory.retain_batch_async(
            bank_id=bank_id,
            contents=[
@@ -1,148 +0,0 @@
-"""
-Tests for XPC error recovery in LocalSTCrossEncoder.
-
-This tests the automatic reinitialization of the cross-encoder model when
-XPC connection errors occur on macOS (common in long-running daemon processes).
-"""
-
-import asyncio
-from unittest.mock import MagicMock, patch
-
-import pytest
-
-from hindsight_api.engine.cross_encoder import LocalSTCrossEncoder
-
-
-class TestCrossEncoderXPCErrorRecovery:
-    """Tests for XPC error detection and recovery in LocalSTCrossEncoder."""
-
-    @pytest.fixture
-    def cross_encoder(self):
-        """Create a LocalSTCrossEncoder instance."""
-        return LocalSTCrossEncoder(model_name="cross-encoder/ms-marco-TinyBERT-L-2-v2")
-
-    def test_is_xpc_error_detection(self, cross_encoder):
-        """Test that XPC errors are correctly detected."""
-        # Test various XPC error message formats
-        xpc_error = Exception("Compiler encountered XPC_ERROR_CONNECTION_INVALID (is the OS shutting down?)")
-        assert cross_encoder._is_xpc_error(xpc_error)
-
-        xpc_error2 = Exception("XPC error occurred")
-        assert cross_encoder._is_xpc_error(xpc_error2)
-
-        # Test that non-XPC errors are not detected
-        normal_error = Exception("Some other error")
-        assert not cross_encoder._is_xpc_error(normal_error)
-
-    @pytest.mark.asyncio
-    async def test_predict_with_xpc_recovery(self, cross_encoder):
-        """Test that predict() recovers from XPC errors by reinitializing."""
-        # Initialize the cross-encoder
-        await cross_encoder.initialize()
-
-        # Track calls to reinitialize
-        reinit_called = False
-        original_reinit = cross_encoder._reinitialize_model_sync
-
-        def track_reinit():
-            nonlocal reinit_called
-            reinit_called = True
-            original_reinit()
-
-        # Track predict attempts
-        predict_attempts = []
-        original_predict = cross_encoder._model.predict
-
-        def mock_predict(*args, **kwargs):
-            predict_attempts.append(1)
-            # Only fail on first attempt
-            if len(predict_attempts) == 1:
-                raise RuntimeError("Compiler encountered XPC_ERROR_CONNECTION_INVALID (is the OS shutting down?)")
-            else:
-                # After reinit: succeed
-                return original_predict(*args, **kwargs)
-
-        # Mock the initial predict to fail, reinit happens, then new model succeeds
-        with patch.object(cross_encoder, "_reinitialize_model_sync", side_effect=track_reinit):
-            with patch.object(cross_encoder._model, "predict", side_effect=mock_predict):
-                # This should trigger XPC error on first attempt, then recover and succeed
-                result = await cross_encoder.predict([("query", "document")])
-
-                # Verify we got a result
-                assert result is not None
-                assert len(result) == 1
-                assert isinstance(result[0], float)
-                assert reinit_called  # Should have reinitialized
-                assert len(predict_attempts) >= 1  # At least one attempt was made
-
-    @pytest.mark.asyncio
-    async def test_predict_fails_on_non_xpc_error(self, cross_encoder):
-        """Test that predict() does not retry for non-XPC errors."""
-        # Initialize the cross-encoder
-        await cross_encoder.initialize()
-
-        # Create a mock that raises a non-XPC error
-        def mock_predict(*args, **kwargs):
-            raise RuntimeError("Some other error")
-
-        # Patch the model's predict method
-        with patch.object(cross_encoder._model, "predict", side_effect=mock_predict):
-            # This should fail without retry
-            with pytest.raises(RuntimeError) as exc_info:
-                await cross_encoder.predict([("query", "document")])
-
-            assert "Some other error" in str(exc_info.value)
-
-    @pytest.mark.asyncio
-    async def test_reinitialize_clears_model(self, cross_encoder):
-        """Test that _reinitialize_model_sync properly clears and reinits the model."""
-        # Initialize the cross-encoder
-        await cross_encoder.initialize()
-
-        original_model = cross_encoder._model
-        assert original_model is not None
-
-        # Reinitialize
-        cross_encoder._reinitialize_model_sync()
-
-        # Model should be reinitialized (new instance)
-        assert cross_encoder._model is not None
-        assert cross_encoder._model is not original_model
-
-        # Should still work
-        result = await cross_encoder.predict([("test query", "test document")])
-        assert len(result) == 1
-        assert isinstance(result[0], float)
-
-    @pytest.mark.asyncio
-    async def test_xpc_recovery_exhausts_retries(self, cross_encoder):
-        """Test that XPC recovery gives up after max retries."""
-        # Initialize the cross-encoder
-        await cross_encoder.initialize()
-
-        # Track reinit calls
-        reinit_count = 0
-        original_reinit = cross_encoder._reinitialize_model_sync
-
-        def track_and_fail_reinit():
-            nonlocal reinit_count
-            reinit_count += 1
-            # Call original reinit, but the new model will also be mocked to fail
-            original_reinit()
-            # After reinit, patch the new model too
-            cross_encoder._model.predict = MagicMock(
-                side_effect=RuntimeError("Compiler encountered XPC_ERROR_CONNECTION_INVALID")
-            )
-
-        # Mock that always raises XPC error
-        cross_encoder._model.predict = MagicMock(
-            side_effect=RuntimeError("Compiler encountered XPC_ERROR_CONNECTION_INVALID")
-        )
-
-        with patch.object(cross_encoder, "_reinitialize_model_sync", side_effect=track_and_fail_reinit):
-            # Should try once, reinitialize, try again, and fail
-            with pytest.raises(Exception) as exc_info:
-                await cross_encoder.predict([("query", "document")])
-
-            assert "XPC_ERROR_CONNECTION_INVALID" in str(exc_info.value) or "Failed to recover" in str(exc_info.value)
-            assert reinit_count == 1  # Should have tried to reinitialize once
@@ -9,18 +9,17 @@ Includes tests for:

 import asyncio
 import os
-from datetime import datetime
-
 import pytest
+from datetime import datetime
 from sqlalchemy import create_engine, text

 from hindsight_api import MemoryEngine, RequestContext
-from hindsight_api.engine.cross_encoder import CohereCrossEncoder, LocalSTCrossEncoder
-from hindsight_api.engine.embeddings import CohereEmbeddings, LocalSTEmbeddings, OpenAIEmbeddings
+from hindsight_api.engine.embeddings import LocalSTEmbeddings, OpenAIEmbeddings, CohereEmbeddings
+from hindsight_api.engine.cross_encoder import LocalSTCrossEncoder, CohereCrossEncoder
 from hindsight_api.engine.query_analyzer import DateparserQueryAnalyzer
-from hindsight_api.engine.task_backend import SyncTaskBackend
-from hindsight_api.extensions import TenantContext, TenantExtension
-from hindsight_api.migrations import ensure_embedding_dimension, run_migrations
+from hindsight_api.extensions import TenantExtension, TenantContext
+from hindsight_api.migrations import run_migrations, ensure_embedding_dimension
+

 # =============================================================================
 # Shared Utilities
@@ -36,11 +35,6 @@ class SchemaTenantExtension(TenantExtension):
    async def authenticate(self, request_context: RequestContext) -> TenantContext:
        return TenantContext(schema_name=self.schema_name)

-    async def list_tenants(self) -> list:
-        from hindsight_api.extensions.tenant import Tenant
-
-        return [Tenant(schema=self.schema_name)]
-

 def get_test_schema(prefix: str, worker_id: str) -> str:
    """Get unique schema name per xdist worker."""
@@ -329,7 +323,6 @@ class TestOpenAIEmbeddings:
            pool_max_size=3,
            run_migrations=False,
            tenant_extension=SchemaTenantExtension(schema_name),
-            task_backend=SyncTaskBackend(),
        )

        try:
@@ -399,7 +392,6 @@ class TestOpenAIEmbeddings:
            pool_max_size=3,
            run_migrations=False,
            tenant_extension=SchemaTenantExtension(schema_name),
-            task_backend=SyncTaskBackend(),
        )

        try:
@@ -567,7 +559,6 @@ class TestCohereIntegration:
            pool_max_size=3,
            run_migrations=False,
            tenant_extension=SchemaTenantExtension(schema_name),
-            task_backend=SyncTaskBackend(),
        )

        try:
@@ -1,148 +0,0 @@
-"""
-Tests for XPC error recovery in LocalSTEmbeddings.
-
-This tests the automatic reinitialization of the embedding model when
-XPC connection errors occur on macOS (common in long-running daemon processes).
-"""
-
-import asyncio
-from unittest.mock import MagicMock, patch
-
-import pytest
-
-from hindsight_api.engine.embeddings import LocalSTEmbeddings
-
-
-class TestXPCErrorRecovery:
-    """Tests for XPC error detection and recovery in LocalSTEmbeddings."""
-
-    @pytest.fixture
-    def embeddings(self):
-        """Create a LocalSTEmbeddings instance."""
-        return LocalSTEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
-
-    def test_is_xpc_error_detection(self, embeddings):
-        """Test that XPC errors are correctly detected."""
-        # Test various XPC error message formats
-        xpc_error = Exception("Compiler encountered XPC_ERROR_CONNECTION_INVALID (is the OS shutting down?)")
-        assert embeddings._is_xpc_error(xpc_error)
-
-        xpc_error2 = Exception("XPC error occurred")
-        assert embeddings._is_xpc_error(xpc_error2)
-
-        # Test that non-XPC errors are not detected
-        normal_error = Exception("Some other error")
-        assert not embeddings._is_xpc_error(normal_error)
-
-    @pytest.mark.asyncio
-    async def test_encode_with_xpc_recovery(self, embeddings):
-        """Test that encode() recovers from XPC errors by reinitializing."""
-        # Initialize the embeddings
-        await embeddings.initialize()
-
-        # Track calls to reinitialize
-        reinit_called = False
-        original_reinit = embeddings._reinitialize_model_sync
-
-        def track_reinit():
-            nonlocal reinit_called
-            reinit_called = True
-            original_reinit()
-
-        # Track encode attempts
-        encode_attempts = []
-        original_encode = embeddings._model.encode
-
-        def mock_encode(*args, **kwargs):
-            encode_attempts.append(1)
-            # Only fail on first attempt
-            if len(encode_attempts) == 1:
-                raise RuntimeError("Compiler encountered XPC_ERROR_CONNECTION_INVALID (is the OS shutting down?)")
-            else:
-                # After reinit: succeed
-                return original_encode(*args, **kwargs)
-
-        # Mock the initial encode to fail, reinit happens, then new model succeeds
-        with patch.object(embeddings, "_reinitialize_model_sync", side_effect=track_reinit):
-            with patch.object(embeddings._model, "encode", side_effect=mock_encode):
-                # This should trigger XPC error on first attempt, then recover and succeed
-                result = embeddings.encode(["test text"])
-
-                # Verify we got a result
-                assert result is not None
-                assert len(result) == 1
-                assert len(result[0]) > 0  # Should have embedding vector
-                assert reinit_called  # Should have reinitialized
-                assert len(encode_attempts) >= 1  # At least one attempt was made
-
-    @pytest.mark.asyncio
-    async def test_encode_fails_on_non_xpc_error(self, embeddings):
-        """Test that encode() does not retry for non-XPC errors."""
-        # Initialize the embeddings
-        await embeddings.initialize()
-
-        # Create a mock that raises a non-XPC error
-        def mock_encode(*args, **kwargs):
-            raise RuntimeError("Some other error")
-
-        # Patch the model's encode method
-        with patch.object(embeddings._model, "encode", side_effect=mock_encode):
-            # This should fail without retry
-            with pytest.raises(RuntimeError) as exc_info:
-                embeddings.encode(["test text"])
-
-            assert "Some other error" in str(exc_info.value)
-
-    @pytest.mark.asyncio
-    async def test_reinitialize_clears_model(self, embeddings):
-        """Test that _reinitialize_model_sync properly clears and reinits the model."""
-        # Initialize the embeddings
-        await embeddings.initialize()
-
-        original_model = embeddings._model
-        assert original_model is not None
-
-        # Reinitialize
-        embeddings._reinitialize_model_sync()
-
-        # Model should be reinitialized (new instance)
-        assert embeddings._model is not None
-        assert embeddings._model is not original_model
-
-        # Should still work
-        result = embeddings.encode(["test"])
-        assert len(result) == 1
-        assert len(result[0]) > 0
-
-    @pytest.mark.asyncio
-    async def test_xpc_recovery_exhausts_retries(self, embeddings):
-        """Test that XPC recovery gives up after max retries."""
-        # Initialize the embeddings
-        await embeddings.initialize()
-
-        # Track reinit calls
-        reinit_count = 0
-        original_reinit = embeddings._reinitialize_model_sync
-
-        def track_and_fail_reinit():
-            nonlocal reinit_count
-            reinit_count += 1
-            # Call original reinit, but the new model will also be mocked to fail
-            original_reinit()
-            # After reinit, patch the new model too
-            embeddings._model.encode = MagicMock(
-                side_effect=RuntimeError("Compiler encountered XPC_ERROR_CONNECTION_INVALID")
-            )
-
-        # Mock that always raises XPC error
-        embeddings._model.encode = MagicMock(
-            side_effect=RuntimeError("Compiler encountered XPC_ERROR_CONNECTION_INVALID")
-        )
-
-        with patch.object(embeddings, "_reinitialize_model_sync", side_effect=track_and_fail_reinit):
-            # Should try once, reinitialize, try again, and fail
-            with pytest.raises(RuntimeError) as exc_info:
-                embeddings.encode(["test"])
-
-            assert "XPC_ERROR_CONNECTION_INVALID" in str(exc_info.value)
-            assert reinit_count == 1  # Should have tried to reinitialize once
@@ -24,9 +24,6 @@ from hindsight_api.extensions import (
    TenantExtension,
    ValidationResult,
    load_extension,
-    # Consolidation operation
-    ConsolidateContext,
-    ConsolidateResult,
 )


@@ -131,18 +128,14 @@ class TrackingValidator(OperationValidatorExtension):

    def __init__(self, config: dict):
        super().__init__(config)
-        # Pre-hook tracking - Core operations
+        # Pre-hook tracking
        self.pre_retain_calls: list[RetainContext] = []
        self.pre_recall_calls: list[RecallContext] = []
        self.pre_reflect_calls: list[ReflectContext] = []
-        # Post-hook tracking - Core operations
+        # Post-hook tracking
        self.post_retain_calls: list[RetainResult] = []
        self.post_recall_calls: list[RecallResult] = []
        self.post_reflect_calls: list[ReflectResultContext] = []
-        # Pre-hook tracking - Consolidation
-        self.pre_consolidate_calls: list[ConsolidateContext] = []
-        # Post-hook tracking - Consolidation
-        self.post_consolidate_calls: list[ConsolidateResult] = []

    async def validate_retain(self, ctx: RetainContext) -> ValidationResult:
        self.pre_retain_calls.append(ctx)
@@ -165,14 +158,6 @@ class TrackingValidator(OperationValidatorExtension):
    async def on_reflect_complete(self, result: ReflectResultContext) -> None:
        self.post_reflect_calls.append(result)

-    # Consolidation hooks
-    async def validate_consolidate(self, ctx: ConsolidateContext) -> ValidationResult:
-        self.pre_consolidate_calls.append(ctx)
-        return ValidationResult.accept()
-
-    async def on_consolidate_complete(self, result: ConsolidateResult) -> None:
-        self.post_consolidate_calls.append(result)
-

 class TestMemoryEngineValidation:
    """Tests for validation integration with MemoryEngine.
@@ -947,3 +947,172 @@ so the algorithm learns to box out. See you next week!
                    raise e


+# =============================================================================
+# DISPOSITION INFERENCE TESTS
+# =============================================================================
+
+class TestDispositionInference:
+    """Tests for LLM-based disposition trait inference from background."""
+
+    @pytest.mark.asyncio
+    async def test_background_merge_with_disposition_inference(self, memory, request_context):
+        """Test that background merge infers disposition traits by default."""
+        import uuid
+        bank_id = f"test_infer_{uuid.uuid4().hex[:8]}"
+
+        result = await memory.merge_bank_background(
+            bank_id,
+            "I am a creative software engineer who loves innovation and trying new technologies",
+            update_disposition=True,
+            request_context=request_context,
+        )
+
+        assert "background" in result
+        assert "disposition" in result
+
+        background = result["background"]
+        disposition = result["disposition"]
+
+        assert "creative" in background.lower() or "innovation" in background.lower()
+
+        # Check that new traits are present with valid values (1-5)
+        required_traits = ["skepticism", "literalism", "empathy"]
+        for trait in required_traits:
+            assert trait in disposition
+            assert 1 <= disposition[trait] <= 5
+
+    @pytest.mark.asyncio
+    async def test_background_merge_without_disposition_inference(self, memory, request_context):
+        """Test that background merge skips disposition inference when disabled."""
+        import uuid
+        bank_id = f"test_no_infer_{uuid.uuid4().hex[:8]}"
+
+        initial_profile = await memory.get_bank_profile(bank_id, request_context=request_context)
+        initial_disposition = initial_profile["disposition"]
+
+        result = await memory.merge_bank_background(
+            bank_id,
+            "I am a data scientist",
+            update_disposition=False,
+            request_context=request_context,
+        )
+
+        assert "background" in result
+        assert "disposition" not in result
+
+        final_profile = await memory.get_bank_profile(bank_id, request_context=request_context)
+        final_disposition = final_profile["disposition"]
+
+        assert initial_disposition == final_disposition
+
+    @pytest.mark.asyncio
+    async def test_disposition_inference_for_lawyer(self, memory, request_context):
+        """Test disposition inference for lawyer profile (high skepticism, high literalism)."""
+        import uuid
+        bank_id = f"test_lawyer_{uuid.uuid4().hex[:8]}"
+
+        result = await memory.merge_bank_background(
+            bank_id,
+            "I am a lawyer who focuses on contract details and never takes claims at face value",
+            update_disposition=True,
+            request_context=request_context,
+        )
+
+        disposition = result["disposition"]
+
+        # Lawyers should have higher skepticism and literalism
+        assert disposition["skepticism"] >= 3
+        assert disposition["literalism"] >= 3
+
+    @pytest.mark.asyncio
+    async def test_disposition_inference_for_therapist(self, memory, request_context):
+        """Test disposition inference for therapist profile (high empathy)."""
+        import uuid
+        bank_id = f"test_therapist_{uuid.uuid4().hex[:8]}"
+
+        result = await memory.merge_bank_background(
+            bank_id,
+            "I am a therapist who deeply understands and connects with people's emotional struggles",
+            update_disposition=True,
+            request_context=request_context,
+        )
+
+        disposition = result["disposition"]
+
+        # Therapists should have higher empathy
+        assert disposition["empathy"] >= 3
+
+    @pytest.mark.asyncio
+    async def test_disposition_updates_in_database(self, memory, request_context):
+        """Test that inferred disposition is actually stored in database."""
+        import uuid
+        bank_id = f"test_db_update_{uuid.uuid4().hex[:8]}"
+
+        result = await memory.merge_bank_background(
+            bank_id,
+            "I am an innovative designer",
+            update_disposition=True,
+            request_context=request_context,
+        )
+
+        inferred_disposition = result["disposition"]
+
+        profile = await memory.get_bank_profile(bank_id, request_context=request_context)
+        db_disposition = profile["disposition"]
+
+        # Compare values (db_disposition is a Pydantic model)
+        assert db_disposition.skepticism == inferred_disposition["skepticism"]
+        assert db_disposition.literalism == inferred_disposition["literalism"]
+        assert db_disposition.empathy == inferred_disposition["empathy"]
+
+    @pytest.mark.asyncio
+    async def test_multiple_background_merges_update_disposition(self, memory, request_context):
+        """Test that each background merge can update disposition."""
+        import uuid
+        bank_id = f"test_multi_merge_{uuid.uuid4().hex[:8]}"
+
+        result1 = await memory.merge_bank_background(
+            bank_id,
+            "I am a software engineer",
+            update_disposition=True,
+            request_context=request_context,
+        )
+        disposition1 = result1["disposition"]
+
+        result2 = await memory.merge_bank_background(
+            bank_id,
+            "I love creative problem solving and innovation",
+            update_disposition=True,
+            request_context=request_context,
+        )
+        disposition2 = result2["disposition"]
+
+        assert "engineer" in result2["background"].lower() or "software" in result2["background"].lower()
+        assert "creative" in result2["background"].lower() or "innovation" in result2["background"].lower()
+
+    @pytest.mark.asyncio
+    async def test_background_merge_conflict_resolution_with_disposition(self, memory, request_context):
+        """Test that conflicts are resolved and disposition reflects final background."""
+        import uuid
+        bank_id = f"test_conflict_{uuid.uuid4().hex[:8]}"
+
+        await memory.merge_bank_background(
+            bank_id,
+            "I was born in Colorado and prefer stability",
+            update_disposition=True,
+            request_context=request_context,
+        )
+
+        result = await memory.merge_bank_background(
+            bank_id,
+            "You were born in Texas and are very skeptical of people",
+            update_disposition=True,
+            request_context=request_context,
+        )
+
+        background = result["background"]
+        disposition = result["disposition"]
+
+        assert "texas" in background.lower()
+        # Higher skepticism expected from "very skeptical of people"
+        assert disposition["skepticism"] >= 3
@@ -51,7 +51,7 @@ Marcus: Yeah, I realized I was being too optimistic about their defense.
    results = await memory.recall_async(
        bank_id=bank_id,
        query="Marcus prediction Rams",
-        fact_type=['experience', 'world'],
+        fact_type=['opinion', 'experience', 'world'],
        budget=Budget.LOW,
        max_tokens=8192,
        request_context=request_context,
@@ -61,8 +61,8 @@ Marcus: Yeah, I realized I was being too optimistic about their defense.
    for i, result in enumerate(results.results):
        print(f"{i+1}. [{result.mentioned_at}] {result.text[:100]}")

-    # Get all facts (Marcus's predictions/statements)
-    agent_facts = results.results
+    # Get all opinion facts (Marcus's predictions/statements)
+    agent_facts = [r for r in results.results if r.fact_type == 'opinion']

    print(f"\n=== Agent facts (Marcus's statements) ===")
    for i, fact in enumerate(agent_facts):
@@ -70,7 +70,6 @@ Marcus: Yeah, I realized I was being too optimistic about their defense.

    # Check that agent facts have different timestamps
    if len(agent_facts) >= 2:
-        # Parse timestamps
        timestamps = [datetime.fromisoformat(f.mentioned_at.replace('Z', '+00:00')) for f in agent_facts]

        # Verify timestamps are different (have time offsets)
@@ -78,40 +77,42 @@ Marcus: Yeah, I realized I was being too optimistic about their defense.
        assert len(unique_timestamps) == len(timestamps), \
            f"Expected unique timestamps for each fact, but got duplicates: {timestamps}"

-        # Sort facts by timestamp for ordering check
-        # Note: recall returns by relevance, not time order
-        sorted_facts = sorted(agent_facts, key=lambda f: datetime.fromisoformat(f.mentioned_at.replace('Z', '+00:00')))
-        sorted_timestamps = [datetime.fromisoformat(f.mentioned_at.replace('Z', '+00:00')) for f in sorted_facts]
-
-        # Verify sorted timestamps are in ascending order
-        for i in range(len(sorted_timestamps) - 1):
-            assert sorted_timestamps[i] < sorted_timestamps[i + 1], \
-                f"Facts should have sequential timestamps. Fact {i} ({sorted_timestamps[i]}) >= Fact {i+1} ({sorted_timestamps[i+1]})"
+        # Verify timestamps are in order (ascending)
+        for i in range(len(timestamps) - 1):
+            assert timestamps[i] < timestamps[i + 1], \
+                f"Facts should be ordered by time. Fact {i} ({timestamps[i]}) >= Fact {i+1} ({timestamps[i+1]})"

        # Verify reasonable time spacing (should be ~10 seconds apart)
-        time_diffs = [(sorted_timestamps[i+1] - sorted_timestamps[i]).total_seconds() for i in range(len(sorted_timestamps) - 1)]
+        time_diffs = [(timestamps[i+1] - timestamps[i]).total_seconds() for i in range(len(timestamps) - 1)]
        print(f"\n=== Time differences between facts: {time_diffs} seconds ===")

        # Each fact should be 10+ seconds apart (allowing for some flexibility)
        for diff in time_diffs:
            assert diff >= 5, f"Expected at least 5 seconds between facts, got {diff}"

-        # Update agent_facts to be sorted for subsequent checks
-        agent_facts = sorted_facts
-        timestamps = sorted_timestamps
-
        print(f"\n✅ All {len(agent_facts)} agent facts have properly ordered timestamps")

-    # Verify that facts capture the key information
-    # Note: LLM may merge related predictions into single facts
+    # Verify that retrieval returns facts in chronological order
+    # The first prediction should come before the changed prediction
    agent_texts = [f.text.lower() for f in agent_facts]
-    all_text = " ".join(agent_texts)

-    # Look for evidence of the predictions being captured (may be merged or separate)
-    has_prediction_info = '27' in all_text or 'rams' in all_text or 'prediction' in all_text
+    # Look for evidence of the sequence
+    has_first_prediction = any('27' in text and '24' in text for text in agent_texts)
+    has_changed_prediction = any('chang' in text or 'by 3' in text or 'realized' in text for text in agent_texts)

-    assert has_prediction_info, "Facts should contain information about Marcus's predictions"
-    print(f"\n✅ Facts capture prediction information")
+    if has_first_prediction and has_changed_prediction:
+        # Find indices
+        first_idx = next(i for i, text in enumerate(agent_texts) if '27' in text and '24' in text)
+        changed_idx = next(i for i, text in enumerate(agent_texts) if 'chang' in text or 'by 3' in text or 'realized' in text)
+
+        print(f"\nFirst prediction at index {first_idx}: {agent_facts[first_idx].text[:100]}")
+        print(f"Changed prediction at index {changed_idx}: {agent_facts[changed_idx].text[:100]}")
+
+        # The original prediction should come before the changed one
+        assert timestamps[first_idx] < timestamps[changed_idx], \
+            "Original prediction should have earlier timestamp than changed prediction"
+
+        print(f"\n✅ Temporal ordering preserved: First prediction came before changed prediction")

    # Cleanup
    await memory.delete_bank(bank_id, request_context=request_context)
@@ -155,14 +156,14 @@ Alice: I reconsidered the team's experience level.
    results = await memory.recall_async(
        bank_id=bank_id,
        query="Alice preference React Vue",
-        fact_type=['experience', 'world'],
+        fact_type=['opinion', 'experience'],
        budget=Budget.LOW,
        max_tokens=8192,
        request_context=request_context,
    )

    print(f"\n=== Retrieved {len(results.results)} agent facts ===")
-    agent_facts = results.results
+    agent_facts = [r for r in results.results if r.fact_type in ('opinion', 'experience')]

    for i, fact in enumerate(agent_facts):
        print(f"{i+1}. [{fact.mentioned_at}] {fact.text[:80]}")
@@ -60,6 +60,17 @@ async def test_full_api_workflow(api_client, test_bank_id):
    assert response.status_code == 200
    profile = response.json()
    assert "disposition" in profile
+    assert "background" in profile
+
+    # Add background
+    response = await api_client.post(
+        f"/v1/default/banks/{test_bank_id}/background",
+        json={
+            "content": "A software engineer passionate about AI and memory systems."
+        }
+    )
+    assert response.status_code == 200
+    assert "software engineer" in response.json()["background"].lower()

    # ================================================================
    # 2. Memory Storage
@@ -233,9 +244,7 @@ async def test_full_api_workflow(api_client, test_bank_id):
    response = await api_client.get(f"/v1/default/banks/{test_bank_id}/profile")
    assert response.status_code == 200
    updated_profile = response.json()
-    assert updated_profile["disposition"]["skepticism"] == 4
-    assert updated_profile["disposition"]["literalism"] == 3
-    assert updated_profile["disposition"]["empathy"] == 4
+    assert "software engineer" in updated_profile["background"].lower()

    # ================================================================
    # 8. Test Entity Endpoints
@@ -280,11 +289,11 @@ async def test_full_api_workflow(api_client, test_bank_id):
        entity_detail = response.json()
        assert "id" in entity_detail

-        # Test regenerate observations (deprecated - returns 410 Gone)
+        # Test regenerate observations
        response = await api_client.post(
            f"/v1/default/banks/{test_bank_id}/entities/{entity_id}/regenerate"
        )
-        assert response.status_code == 410  # Deprecated endpoint
+        assert response.status_code == 200

    # ================================================================
    # 9. List All Banks (should include our test bank)
@@ -836,8 +845,9 @@ async def test_reflect_structured_output(api_client):
    assert response.status_code == 200
    result = response.json()

-    # Verify text field exists (may contain text even with structured output)
+    # Verify text field exists (empty when using structured output)
    assert "text" in result
+    assert result["text"] == ""

    # Verify structured output exists and has expected structure
    assert "structured_output" in result
@@ -969,12 +979,10 @@ async def test_reflect_returns_token_usage(api_client):
    assert "text" in result
    assert len(result["text"]) > 0

-    # Verify usage field exists and is populated (agentic reflect aggregates all LLM calls)
+    # Verify usage field exists and has expected structure
    assert "usage" in result, "Response should include 'usage' field"
    usage = result["usage"]
-
-    # Usage must be present - agentic reflect now aggregates token usage from all LLM calls
-    assert usage is not None, "Usage should not be None - reflect aggregates all LLM call usages"
+    assert usage is not None, "Usage should not be None for reflect"
    assert "input_tokens" in usage, "Usage should have 'input_tokens'"
    assert "output_tokens" in usage, "Usage should have 'output_tokens'"
    assert "total_tokens" in usage, "Usage should have 'total_tokens'"
@@ -1,278 +0,0 @@
-"""
-Tests for LinkExpansion graph retrieval.
-
-Tests cover the entity-based graph traversal for observations.
-"""
-
-from datetime import datetime, timezone
-
-import pytest
-
-
-@pytest.fixture(autouse=True)
-def enable_observations():
-    """Enable observations for all tests in this module."""
-    from hindsight_api.config import get_config
-
-    config = get_config()
-    original_value = config.enable_observations
-    config.enable_observations = True
-    yield
-    config.enable_observations = original_value
-
-
-@pytest.mark.asyncio
-async def test_link_expansion_observation_graph_retrieval(memory, request_context):
-    """
-    Test that observations can find other observations via shared entities.
-
-    This tests the scenario where:
-    1. World fact A has entity "Python"
-    2. World fact B has entity "Python"
-    3. Observation OA is derived from world fact A
-    4. Observation OB is derived from world fact B
-
-    When searching for observations related to OA, graph retrieval should find OB
-    because they share the "Python" entity through their source world facts.
-
-    Current issue: Graph retrieval returns 0 for observations because:
-    - Entity links are copied from world facts to observations during consolidation
-    - But the entity expansion query filters by fact_type
-    - Observations only share entities with world facts (cross-type), not with other observations
-    - So filtering to fact_type='observation' returns 0 results
-    """
-    bank_id = f"test_link_expansion_obs_{datetime.now(timezone.utc).timestamp()}"
-
-    try:
-        # Store world facts with shared entities using retain_batch_async
-        # We need enough facts that semantic search won't return all of them as seeds
-        # Key: "Alice" query should find Alice's observation but NOT Bob's via semantic search
-        # Then graph retrieval should find Bob via shared "Python" entity
-        await memory.retain_batch_async(
-            bank_id=bank_id,
-            contents=[
-                # Python developers - should be connected via "Python" entity
-                {
-                    "content": "Alice works with Python at TechCorp building REST APIs",
-                    "context": "employee info",
-                    "entities": [{"text": "Python"}, {"text": "Alice"}, {"text": "TechCorp"}],
-                },
-                {
-                    "content": "Bob uses Python at DataSoft for machine learning models",
-                    "context": "employee info",
-                    "entities": [{"text": "Python"}, {"text": "Bob"}, {"text": "DataSoft"}],
-                },
-                # Many unrelated facts to dilute semantic search and ensure
-                # "Alice" query only finds Alice-related content as seeds
-                {
-                    "content": "The weather in San Francisco is often foggy and cool",
-                    "context": "weather info",
-                    "entities": [{"text": "San Francisco"}],
-                },
-                {
-                    "content": "Tokyo is the capital city of Japan with many trains",
-                    "context": "geography info",
-                    "entities": [{"text": "Tokyo"}, {"text": "Japan"}],
-                },
-                {
-                    "content": "The Great Wall of China is a historic fortification",
-                    "context": "history info",
-                    "entities": [{"text": "Great Wall"}, {"text": "China"}],
-                },
-                {
-                    "content": "Coffee beans are grown in tropical regions worldwide",
-                    "context": "food info",
-                    "entities": [{"text": "Coffee"}],
-                },
-                {
-                    "content": "Electric vehicles are becoming more popular globally",
-                    "context": "technology info",
-                    "entities": [{"text": "Electric vehicles"}],
-                },
-                {
-                    "content": "The Amazon rainforest contains diverse wildlife species",
-                    "context": "nature info",
-                    "entities": [{"text": "Amazon"}, {"text": "Rainforest"}],
-                },
-                {
-                    "content": "Basketball is a popular sport in the United States",
-                    "context": "sports info",
-                    "entities": [{"text": "Basketball"}, {"text": "United States"}],
-                },
-                {
-                    "content": "Mozart composed many famous classical music pieces",
-                    "context": "music info",
-                    "entities": [{"text": "Mozart"}, {"text": "Classical music"}],
-                },
-            ],
-            request_context=request_context,
-        )
-
-        # Consolidation runs automatically after retain - wait for it to complete
-        # by querying for observations (consolidation creates them)
-        import asyncio
-        from hindsight_api.engine.memory_engine import Budget
-
-        # Wait for consolidation to complete with retry logic
-        # Consolidation runs as a background task and may take longer in CI
-        obs_result = None
-        for _ in range(30):  # Try up to 30 times (30 seconds max)
-            await asyncio.sleep(1)  # Wait 1 second between attempts
-            obs_result = await memory.recall_async(
-                bank_id=bank_id,
-                query="Python developer",
-                fact_type=["observation"],
-                budget=Budget.MID,
-                max_tokens=2048,
-                request_context=request_context,
-            )
-            if obs_result.results and len(obs_result.results) >= 1:
-                break
-
-        assert obs_result is not None and obs_result.results is not None, "Should have observations after consolidation"
-        # We should have observations from consolidation
-        assert len(obs_result.results) >= 1, f"Should have at least 1 observation about Python, got {len(obs_result.results)}"
-
-        # Now test graph retrieval specifically
-        # Query for Alice - should find Bob via shared "Python" entity
-        result = await memory.recall_async(
-            bank_id=bank_id,
-            query="Alice",
-            fact_type=["observation"],
-            budget=Budget.MID,
-            max_tokens=2048,
-            enable_trace=True,
-            request_context=request_context,
-        )
-
-        # Verify graph retrieval is working by checking the internal debug logs
-        # The graph retrieval finds observations via entity links, but may not return
-        # NEW results if semantic search already found all connected observations.
-        # This is correct behavior - we verify the entity traversal path works.
-
-        # Check the trace for graph results
-        assert result.trace is not None, "Should have trace data"
-
-        # The key verification: the entity expansion path works (sources -> entities -> observations)
-        # We validated this in the debug logs above:
-        # - Observations have source_memory_ids pointing to world facts ✓
-        # - World facts have entity links ✓
-        # - Graph retrieval can traverse this path (seen in logs: potential_obs > 0)
-
-        # For a more rigorous test, we need data where semantic search misses something.
-        # Let's verify the world fact graph retrieval works (it uses direct entity links).
-        world_result = await memory.recall_async(
-            bank_id=bank_id,
-            query="Alice",
-            fact_type=["world"],
-            budget=Budget.MID,
-            max_tokens=2048,
-            enable_trace=True,
-            request_context=request_context,
-        )
-
-        assert world_result.trace is not None, "Should have trace data for world facts"
-        world_retrieval_results = world_result.trace.get("retrieval_results", [])
-        world_graph_results = [
-            r for r in world_retrieval_results if r.get("method_name") == "graph"
-        ]
-
-        if world_graph_results:
-            world_graph_result = [r for r in world_graph_results if r.get("fact_type") == "world"][0]
-            world_graph_results_list = world_graph_result.get("results", [])
-
-            # World facts use direct entity links, so graph may find results
-            if world_graph_results_list:
-                print(f"\n✓ Graph retrieval found {len(world_graph_results_list)} connected world facts")
-                graph_texts = [r.get("text", "") for r in world_graph_results_list]
-                bob_found = any("Bob" in t or "DataSoft" in t for t in graph_texts)
-                if bob_found:
-                    print("  Found Bob's world fact via shared 'Python' entity!")
-
-        print("\n✓ Link expansion observation test passed!")
-        print("  Entity traversal path verified (observations -> sources -> entities -> connected sources -> observations)")
-
-    finally:
-        await memory.delete_bank(bank_id, request_context=request_context)
-
-
-@pytest.mark.asyncio
-async def test_link_expansion_world_fact_graph_retrieval(memory, request_context):
-    """
-    Test that world facts can find other world facts via shared entities.
-
-    This verifies the direct entity link traversal for world facts works correctly.
-    Note: When semantic search finds all world facts as seeds, graph retrieval
-    won't return NEW results (this is correct - it shouldn't duplicate results).
-    """
-    bank_id = f"test_link_expansion_world_{datetime.now(timezone.utc).timestamp()}"
-
-    try:
-        # Store world facts with shared entities
-        await memory.retain_batch_async(
-            bank_id=bank_id,
-            contents=[
-                # Python developers - should be connected via "Python" entity
-                {
-                    "content": "Alice works with Python at TechCorp building REST APIs",
-                    "context": "employee info",
-                    "entities": [{"text": "Python"}, {"text": "Alice"}, {"text": "TechCorp"}],
-                },
-                {
-                    "content": "Bob uses Python at DataSoft for machine learning models",
-                    "context": "employee info",
-                    "entities": [{"text": "Python"}, {"text": "Bob"}, {"text": "DataSoft"}],
-                },
-                # Unrelated facts
-                {
-                    "content": "The weather in San Francisco is often foggy",
-                    "context": "weather info",
-                    "entities": [{"text": "San Francisco"}],
-                },
-                {
-                    "content": "Coffee beans are grown in tropical regions",
-                    "context": "food info",
-                    "entities": [{"text": "Coffee"}],
-                },
-            ],
-            request_context=request_context,
-        )
-
-        from hindsight_api.engine.memory_engine import Budget
-
-        # Query for Alice
-        result = await memory.recall_async(
-            bank_id=bank_id,
-            query="Alice",
-            fact_type=["world"],
-            budget=Budget.MID,
-            max_tokens=2048,
-            enable_trace=True,
-            request_context=request_context,
-        )
-
-        assert result.trace is not None, "Should have trace data"
-
-        # Verify graph retrieval ran (it may or may not find new results depending
-        # on whether semantic search already found everything)
-        retrieval_results = result.trace.get("retrieval_results", [])
-        graph_results = [
-            r for r in retrieval_results if r.get("method_name") == "graph"
-        ]
-        assert len(graph_results) > 0, "Should have graph retrieval results in trace"
-
-        # The important thing is that recall works and returns relevant results
-        assert result.results is not None and len(result.results) > 0, (
-            "Should return results for 'Alice' query"
-        )
-
-        # Alice's result should be at or near the top
-        result_texts = [r.text for r in result.results]
-        alice_found = any("Alice" in t for t in result_texts)
-        assert alice_found, f"Should find Alice in results: {result_texts[:3]}"
-
-        print("\n✓ Link expansion world fact test passed!")
-        print(f"  Recall returned {len(result.results)} results for 'Alice' query")
-
-    finally:
-        await memory.delete_bank(bank_id, request_context=request_context)
@@ -1,329 +0,0 @@
-"""
-Tests for LLM tool calling functionality.
-"""
-
-import pytest
-
-from hindsight_api.engine.llm_wrapper import LLMProvider
-from hindsight_api.engine.response_models import LLMToolCall, LLMToolCallResult
-
-
-# Sample tools for testing
-SAMPLE_TOOLS = [
-    {
-        "type": "function",
-        "function": {
-            "name": "get_weather",
-            "description": "Get weather for a location",
-            "parameters": {
-                "type": "object",
-                "properties": {
-                    "location": {"type": "string", "description": "City name"},
-                    "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]},
-                },
-                "required": ["location"],
-            },
-        },
-    },
-    {
-        "type": "function",
-        "function": {
-            "name": "search",
-            "description": "Search for information",
-            "parameters": {
-                "type": "object",
-                "properties": {
-                    "query": {"type": "string", "description": "Search query"},
-                },
-                "required": ["query"],
-            },
-        },
-    },
-]
-
-
-class TestMockToolCalling:
-    """Test tool calling with mock provider."""
-
-    @pytest.mark.asyncio
-    async def test_call_with_tools_returns_tool_calls(self):
-        """Test that mock provider can return tool calls."""
-        llm = LLMProvider(provider="mock", api_key="", base_url="", model="mock")
-
-        # Set mock response to return tool calls
-        llm.set_mock_response([
-            {"name": "get_weather", "arguments": {"location": "Paris", "unit": "celsius"}},
-        ])
-
-        result = await llm.call_with_tools(
-            messages=[{"role": "user", "content": "What's the weather in Paris?"}],
-            tools=SAMPLE_TOOLS,
-        )
-
-        assert isinstance(result, LLMToolCallResult)
-        assert len(result.tool_calls) == 1
-        assert result.tool_calls[0].name == "get_weather"
-        assert result.tool_calls[0].arguments == {"location": "Paris", "unit": "celsius"}
-        assert result.finish_reason == "tool_calls"
-
-    @pytest.mark.asyncio
-    async def test_call_with_tools_returns_content(self):
-        """Test that mock provider can return plain content."""
-        llm = LLMProvider(provider="mock", api_key="", base_url="", model="mock")
-
-        # Default mock response is plain content
-        result = await llm.call_with_tools(
-            messages=[{"role": "user", "content": "Hello"}],
-            tools=SAMPLE_TOOLS,
-        )
-
-        assert isinstance(result, LLMToolCallResult)
-        assert result.content == "mock response"
-        assert len(result.tool_calls) == 0
-        assert result.finish_reason == "stop"
-
-    @pytest.mark.asyncio
-    async def test_call_with_tools_records_calls(self):
-        """Test that mock calls are recorded."""
-        llm = LLMProvider(provider="mock", api_key="", base_url="", model="mock")
-        llm.clear_mock_calls()
-
-        await llm.call_with_tools(
-            messages=[{"role": "user", "content": "Test message"}],
-            tools=SAMPLE_TOOLS,
-            scope="test_scope",
-        )
-
-        calls = llm.get_mock_calls()
-        assert len(calls) == 1
-        assert calls[0]["scope"] == "test_scope"
-        assert "get_weather" in calls[0]["tools"]
-        assert "search" in calls[0]["tools"]
-
-    @pytest.mark.asyncio
-    async def test_call_with_tools_multiple_tool_calls(self):
-        """Test handling multiple tool calls in one response."""
-        llm = LLMProvider(provider="mock", api_key="", base_url="", model="mock")
-
-        llm.set_mock_response([
-            {"name": "get_weather", "arguments": {"location": "Paris"}},
-            {"name": "search", "arguments": {"query": "weather forecast"}},
-        ])
-
-        result = await llm.call_with_tools(
-            messages=[{"role": "user", "content": "Weather in Paris and search for forecasts"}],
-            tools=SAMPLE_TOOLS,
-        )
-
-        assert len(result.tool_calls) == 2
-        assert result.tool_calls[0].name == "get_weather"
-        assert result.tool_calls[1].name == "search"
-
-    @pytest.mark.asyncio
-    async def test_call_with_tools_accepts_llm_tool_call_result(self):
-        """Test that mock can accept LLMToolCallResult directly."""
-        llm = LLMProvider(provider="mock", api_key="", base_url="", model="mock")
-
-        expected_result = LLMToolCallResult(
-            content="Here's the info",
-            tool_calls=[LLMToolCall(id="call_123", name="search", arguments={"query": "test"})],
-            finish_reason="tool_calls",
-        )
-        llm.set_mock_response(expected_result)
-
-        result = await llm.call_with_tools(
-            messages=[{"role": "user", "content": "Search for test"}],
-            tools=SAMPLE_TOOLS,
-        )
-
-        assert result == expected_result
-
-
-class TestToolCallConversation:
-    """Test tool call conversation flow."""
-
-    @pytest.mark.asyncio
-    async def test_tool_result_message_format(self):
-        """Test that tool result messages can be passed in subsequent calls."""
-        llm = LLMProvider(provider="mock", api_key="", base_url="", model="mock")
-
-        # First call returns tool call
-        llm.set_mock_response([{"name": "get_weather", "arguments": {"location": "Paris"}}])
-
-        result1 = await llm.call_with_tools(
-            messages=[{"role": "user", "content": "What's the weather?"}],
-            tools=SAMPLE_TOOLS,
-        )
-
-        # Build conversation with tool result
-        messages = [
-            {"role": "user", "content": "What's the weather?"},
-            {
-                "role": "assistant",
-                "tool_calls": [
-                    {
-                        "id": result1.tool_calls[0].id,
-                        "type": "function",
-                        "function": {
-                            "name": result1.tool_calls[0].name,
-                            "arguments": '{"location": "Paris"}',
-                        },
-                    }
-                ],
-            },
-            {
-                "role": "tool",
-                "tool_call_id": result1.tool_calls[0].id,
-                "content": '{"temperature": 20, "conditions": "sunny"}',
-            },
-        ]
-
-        # Second call should work with tool result in history
-        llm.set_mock_response(None)  # Reset to default
-        result2 = await llm.call_with_tools(
-            messages=messages,
-            tools=SAMPLE_TOOLS,
-        )
-
-        assert result2.content == "mock response"
-
-
-class TestToolSchemas:
-    """Test tool schema handling."""
-
-    @pytest.mark.asyncio
-    async def test_empty_tools_list(self):
-        """Test calling with empty tools list."""
-        llm = LLMProvider(provider="mock", api_key="", base_url="", model="mock")
-
-        result = await llm.call_with_tools(
-            messages=[{"role": "user", "content": "Hello"}],
-            tools=[],
-        )
-
-        assert result.content == "mock response"
-
-    @pytest.mark.asyncio
-    async def test_tool_with_no_required_params(self):
-        """Test tool with no required parameters."""
-        llm = LLMProvider(provider="mock", api_key="", base_url="", model="mock")
-
-        tools = [
-            {
-                "type": "function",
-                "function": {
-                    "name": "list_items",
-                    "description": "List all items",
-                    "parameters": {"type": "object", "properties": {}, "required": []},
-                },
-            }
-        ]
-
-        llm.set_mock_response([{"name": "list_items", "arguments": {}}])
-
-        result = await llm.call_with_tools(
-            messages=[{"role": "user", "content": "List items"}],
-            tools=tools,
-        )
-
-        assert len(result.tool_calls) == 1
-        assert result.tool_calls[0].name == "list_items"
-        assert result.tool_calls[0].arguments == {}
-
-
-class TestReflectToolSchemas:
-    """Test reflect agent tool schemas."""
-
-    def test_get_reflect_tools_default(self):
-        """Test getting default reflect tools."""
-        from hindsight_api.engine.reflect.tools_schema import get_reflect_tools
-
-        tools = get_reflect_tools()
-
-        tool_names = [t["function"]["name"] for t in tools]
-        assert "search_mental_models" in tool_names
-        assert "search_observations" in tool_names
-        assert "recall" in tool_names
-        assert "expand" in tool_names
-        assert "done" in tool_names
-
-    def test_get_reflect_tools_with_directives(self):
-        """Test getting reflect tools with directive rules."""
-        from hindsight_api.engine.reflect.tools_schema import get_reflect_tools
-
-        tools = get_reflect_tools(directive_rules=["Always respond in French"])
-
-        tool_names = [t["function"]["name"] for t in tools]
-        assert "recall" in tool_names
-        assert "done" in tool_names
-
-        # Done tool should have directive_compliance field when directives are present
-        done_tool = next(t for t in tools if t["function"]["name"] == "done")
-        params = done_tool["function"]["parameters"]["properties"]
-        assert "directive_compliance" in params
-
-    def test_get_reflect_tools_answer_mode(self):
-        """Test getting reflect tools with answer output mode."""
-        from hindsight_api.engine.reflect.tools_schema import get_reflect_tools
-
-        tools = get_reflect_tools()
-
-        done_tool = next(t for t in tools if t["function"]["name"] == "done")
-        params = done_tool["function"]["parameters"]["properties"]
-
-        assert "answer" in params
-        assert "memory_ids" in params
-        assert "observation_ids" in params
-        assert "mental_model_ids" in params
-
-
-class TestLLMToolCallResult:
-    """Test LLMToolCallResult model."""
-
-    def test_tool_call_result_defaults(self):
-        """Test default values for LLMToolCallResult."""
-        result = LLMToolCallResult()
-
-        assert result.content is None
-        assert result.tool_calls == []
-        assert result.finish_reason is None
-
-    def test_tool_call_result_with_content(self):
-        """Test LLMToolCallResult with content."""
-        result = LLMToolCallResult(content="Hello", finish_reason="stop")
-
-        assert result.content == "Hello"
-        assert result.tool_calls == []
-        assert result.finish_reason == "stop"
-
-    def test_tool_call_result_with_tool_calls(self):
-        """Test LLMToolCallResult with tool calls."""
-        result = LLMToolCallResult(
-            tool_calls=[
-                LLMToolCall(id="call_1", name="test_tool", arguments={"arg": "value"}),
-            ],
-            finish_reason="tool_calls",
-        )
-
-        assert result.content is None
-        assert len(result.tool_calls) == 1
-        assert result.tool_calls[0].name == "test_tool"
-        assert result.finish_reason == "tool_calls"
-
-
-class TestLLMToolCall:
-    """Test LLMToolCall model."""
-
-    def test_tool_call_basic(self):
-        """Test basic LLMToolCall creation."""
-        call = LLMToolCall(id="call_123", name="get_weather", arguments={"location": "Paris"})
-
-        assert call.id == "call_123"
-        assert call.name == "get_weather"
-        assert call.arguments == {"location": "Paris"}
-
-    def test_tool_call_empty_arguments(self):
-        """Test LLMToolCall with empty arguments."""
-        call = LLMToolCall(id="call_456", name="list_items", arguments={})
-
-        assert call.arguments == {}
@@ -19,7 +19,6 @@ import pytest_asyncio
 from hindsight_api import MemoryEngine, LLMConfig, LocalSTEmbeddings, RequestContext
 from hindsight_api.engine.cross_encoder import LocalSTCrossEncoder
 from hindsight_api.engine.query_analyzer import DateparserQueryAnalyzer
-from hindsight_api.engine.task_backend import SyncTaskBackend
 from hindsight_api.engine.retain.fact_extraction import FactExtractionResponse, ExtractedFact
 from hindsight_api.engine.llm_wrapper import TokenUsage

@@ -107,7 +106,6 @@ class TestLargeBatchRetain:
            pool_max_size=10,
            run_migrations=False,
            skip_llm_verification=True,  # Skip LLM verification since we're mocking
-            task_backend=SyncTaskBackend(),  # Execute tasks immediately in tests
        )
        await mem.initialize()
        yield mem
@@ -1,401 +0,0 @@
-"""
-Tests for hindsight_api.main module (single-worker code path).
-
-The main.py module is used when running with a single worker:
-    hindsight-api  (or hindsight-api --workers 1)
-
-When workers=1, main.py creates the app directly and passes it to uvicorn.
-These tests ensure that extensions are properly loaded in this code path.
-
-Compare with test_server_module.py which tests the multi-worker path (workers > 1).
-"""
-
-import sys
-from unittest.mock import MagicMock, patch
-
-
-class TestMainModuleExtensionLoading:
-    """Tests that main.py correctly loads extensions when configured via environment."""
-
-    def test_main_loads_tenant_extension_when_configured(self, monkeypatch):
-        """
-        Verify that main.py loads tenant extension from HINDSIGHT_API_TENANT_EXTENSION.
-
-        This ensures extension loading works in the single-worker code path.
-        """
-        # Set up environment to configure a tenant extension
-        monkeypatch.setenv(
-            "HINDSIGHT_API_TENANT_EXTENSION",
-            "tests.test_main_module:MockTenantExtension",
-        )
-        # Ensure single worker mode
-        monkeypatch.setenv("HINDSIGHT_API_WORKERS", "1")
-
-        # Track what extensions were loaded via load_extension
-        loaded_extensions = {}
-
-        # Get the real load_extension function
-        from hindsight_api.extensions.loader import load_extension as real_load_extension
-
-        def tracking_load_extension(name, base_class):
-            """Track calls to load_extension and delegate to original."""
-            result = real_load_extension(name, base_class)
-            loaded_extensions[name] = result
-            return result
-
-        with patch("hindsight_api.main.MemoryEngine") as mock_engine, \
-             patch("hindsight_api.main.create_app") as mock_create_app, \
-             patch("hindsight_api.main.get_config") as mock_get_config, \
-             patch("hindsight_api.main.load_extension", side_effect=tracking_load_extension), \
-             patch("hindsight_api.main.DefaultExtensionContext"), \
-             patch("hindsight_api.main.print_banner"), \
-             patch("uvicorn.run"):  # Don't actually start uvicorn
-
-            mock_config = MagicMock()
-            mock_config.host = "0.0.0.0"
-            mock_config.port = 8888
-            mock_config.log_level = "info"
-            mock_config.mcp_enabled = False
-            mock_config.run_migrations_on_startup = False
-            mock_config.database_url = "postgresql://test:test@localhost/test"
-            mock_get_config.return_value = mock_config
-            mock_engine.return_value = MagicMock()
-            mock_create_app.return_value = MagicMock()
-
-            # Mock sys.argv to simulate CLI invocation
-            with patch.object(sys, 'argv', ['hindsight-api']):
-                from hindsight_api.main import main
-                main()
-
-        # Verify TENANT extension was loaded
-        assert "TENANT" in loaded_extensions, \
-            "main.py did not call load_extension('TENANT', ...) - extensions not loaded!"
-        assert loaded_extensions["TENANT"] is not None, \
-            "load_extension('TENANT', ...) returned None despite env var being set"
-        assert isinstance(loaded_extensions["TENANT"], MockTenantExtension), \
-            f"Expected MockTenantExtension, got {type(loaded_extensions['TENANT'])}"
-
-    def test_main_loads_operation_validator_when_configured(self, monkeypatch):
-        """
-        Verify that main.py loads operation validator from HINDSIGHT_API_OPERATION_VALIDATOR_EXTENSION.
-        """
-        monkeypatch.setenv(
-            "HINDSIGHT_API_OPERATION_VALIDATOR_EXTENSION",
-            "tests.test_main_module:MockOperationValidator",
-        )
-        monkeypatch.setenv("HINDSIGHT_API_WORKERS", "1")
-
-        loaded_extensions = {}
-
-        from hindsight_api.extensions.loader import load_extension as real_load_extension
-
-        def tracking_load_extension(name, base_class):
-            result = real_load_extension(name, base_class)
-            loaded_extensions[name] = result
-            return result
-
-        with patch("hindsight_api.main.MemoryEngine") as mock_engine, \
-             patch("hindsight_api.main.create_app") as mock_create_app, \
-             patch("hindsight_api.main.get_config") as mock_get_config, \
-             patch("hindsight_api.main.load_extension", side_effect=tracking_load_extension), \
-             patch("hindsight_api.main.DefaultExtensionContext"), \
-             patch("hindsight_api.main.print_banner"), \
-             patch("uvicorn.run"):
-
-            mock_config = MagicMock()
-            mock_config.host = "0.0.0.0"
-            mock_config.port = 8888
-            mock_config.log_level = "info"
-            mock_config.mcp_enabled = False
-            mock_config.run_migrations_on_startup = False
-            mock_config.database_url = "postgresql://test:test@localhost/test"
-            mock_get_config.return_value = mock_config
-            mock_engine.return_value = MagicMock()
-            mock_create_app.return_value = MagicMock()
-
-            with patch.object(sys, 'argv', ['hindsight-api']):
-                from hindsight_api.main import main
-                main()
-
-        assert "OPERATION_VALIDATOR" in loaded_extensions, \
-            "main.py did not call load_extension('OPERATION_VALIDATOR', ...)"
-        assert loaded_extensions["OPERATION_VALIDATOR"] is not None
-        assert isinstance(loaded_extensions["OPERATION_VALIDATOR"], MockOperationValidator)
-
-    def test_main_passes_extensions_to_memory_engine(self, monkeypatch):
-        """
-        Verify that main.py passes loaded extensions to MemoryEngine constructor.
-
-        This is the critical test - even if extensions are loaded, they must be
-        passed to MemoryEngine for authentication to work.
-        """
-        monkeypatch.setenv(
-            "HINDSIGHT_API_TENANT_EXTENSION",
-            "tests.test_main_module:MockTenantExtension",
-        )
-        monkeypatch.setenv("HINDSIGHT_API_WORKERS", "1")
-
-        memory_engine_calls = []
-
-        def capture_memory_engine(*args, **kwargs):
-            memory_engine_calls.append({"args": args, "kwargs": kwargs})
-            return MagicMock()
-
-        with patch("hindsight_api.main.MemoryEngine", side_effect=capture_memory_engine), \
-             patch("hindsight_api.main.create_app") as mock_create_app, \
-             patch("hindsight_api.main.get_config") as mock_get_config, \
-             patch("hindsight_api.main.DefaultExtensionContext"), \
-             patch("hindsight_api.main.print_banner"), \
-             patch("uvicorn.run"):
-
-            mock_config = MagicMock()
-            mock_config.host = "0.0.0.0"
-            mock_config.port = 8888
-            mock_config.log_level = "info"
-            mock_config.mcp_enabled = False
-            mock_config.run_migrations_on_startup = False
-            mock_config.database_url = "postgresql://test:test@localhost/test"
-            mock_get_config.return_value = mock_config
-            mock_create_app.return_value = MagicMock()
-
-            with patch.object(sys, 'argv', ['hindsight-api']):
-                from hindsight_api.main import main
-                main()
-
-        # Verify MemoryEngine was called
-        assert len(memory_engine_calls) == 1, "MemoryEngine should be called exactly once"
-
-        call_kwargs = memory_engine_calls[0]["kwargs"]
-
-        # THE CRITICAL ASSERTION: tenant_extension must be passed and not None
-        assert "tenant_extension" in call_kwargs, \
-            "MemoryEngine was not called with tenant_extension parameter!"
-        assert call_kwargs["tenant_extension"] is not None, \
-            "tenant_extension was None - main.py did not pass loaded extension to MemoryEngine!"
-
-    def test_main_sets_extension_context_on_tenant_extension(self, monkeypatch):
-        """
-        Verify that main.py sets the extension context on tenant extension.
-
-        This is required for tenant extensions that need to provision schemas.
-        """
-        monkeypatch.setenv(
-            "HINDSIGHT_API_TENANT_EXTENSION",
-            "tests.test_main_module:MockTenantExtension",
-        )
-        monkeypatch.setenv("HINDSIGHT_API_WORKERS", "1")
-
-        captured_tenant_ext = [None]
-
-        def capture_memory_engine(*args, **kwargs):
-            captured_tenant_ext[0] = kwargs.get("tenant_extension")
-            return MagicMock()
-
-        context_created = []
-
-        def capture_context(*args, **kwargs):
-            ctx = MagicMock()
-            context_created.append(ctx)
-            return ctx
-
-        with patch("hindsight_api.main.MemoryEngine", side_effect=capture_memory_engine), \
-             patch("hindsight_api.main.create_app") as mock_create_app, \
-             patch("hindsight_api.main.get_config") as mock_get_config, \
-             patch("hindsight_api.main.DefaultExtensionContext", side_effect=capture_context), \
-             patch("hindsight_api.main.print_banner"), \
-             patch("uvicorn.run"):
-
-            mock_config = MagicMock()
-            mock_config.host = "0.0.0.0"
-            mock_config.port = 8888
-            mock_config.log_level = "info"
-            mock_config.mcp_enabled = False
-            mock_config.run_migrations_on_startup = False
-            mock_config.database_url = "postgresql://test:test@localhost/test"
-            mock_get_config.return_value = mock_config
-            mock_create_app.return_value = MagicMock()
-
-            with patch.object(sys, 'argv', ['hindsight-api']):
-                from hindsight_api.main import main
-                main()
-
-        # Verify context was created and set
-        assert len(context_created) == 1, "DefaultExtensionContext should be created"
-        assert captured_tenant_ext[0] is not None, "Tenant extension should be captured"
-        assert captured_tenant_ext[0]._context_set, \
-            "set_context was not called on tenant extension"
-
-    def test_main_works_without_extensions(self, monkeypatch):
-        """
-        Verify that main.py works correctly when no extensions are configured.
-        """
-        # Ensure no extension env vars are set
-        monkeypatch.delenv("HINDSIGHT_API_TENANT_EXTENSION", raising=False)
-        monkeypatch.delenv("HINDSIGHT_API_OPERATION_VALIDATOR_EXTENSION", raising=False)
-        monkeypatch.setenv("HINDSIGHT_API_WORKERS", "1")
-
-        memory_engine_calls = []
-
-        def capture_memory_engine(*args, **kwargs):
-            memory_engine_calls.append({"args": args, "kwargs": kwargs})
-            return MagicMock()
-
-        with patch("hindsight_api.main.MemoryEngine", side_effect=capture_memory_engine), \
-             patch("hindsight_api.main.create_app") as mock_create_app, \
-             patch("hindsight_api.main.get_config") as mock_get_config, \
-             patch("hindsight_api.main.print_banner"), \
-             patch("uvicorn.run"):
-
-            mock_config = MagicMock()
-            mock_config.host = "0.0.0.0"
-            mock_config.port = 8888
-            mock_config.log_level = "info"
-            mock_config.mcp_enabled = False
-            mock_config.run_migrations_on_startup = False
-            mock_config.database_url = "postgresql://test:test@localhost/test"
-            mock_get_config.return_value = mock_config
-            mock_create_app.return_value = MagicMock()
-
-            with patch.object(sys, 'argv', ['hindsight-api']):
-                from hindsight_api.main import main
-                main()
-
-        # Should work without extensions
-        assert len(memory_engine_calls) == 1
-        call_kwargs = memory_engine_calls[0]["kwargs"]
-
-        # Extensions should be None when not configured
-        assert call_kwargs.get("tenant_extension") is None
-        assert call_kwargs.get("operation_validator") is None
-
-    def test_main_uses_app_object_for_single_worker(self, monkeypatch):
-        """
-        Verify that main.py passes the app object (not import string) when workers=1.
-
-        This is important because it means single-worker mode uses the app created
-        in main.py (with extensions loaded), not server.py.
-        """
-        monkeypatch.setenv("HINDSIGHT_API_WORKERS", "1")
-        monkeypatch.delenv("HINDSIGHT_API_TENANT_EXTENSION", raising=False)
-
-        uvicorn_calls = []
-
-        def capture_uvicorn_run(**kwargs):
-            uvicorn_calls.append(kwargs)
-
-        mock_app = MagicMock()
-
-        with patch("hindsight_api.main.MemoryEngine") as mock_engine, \
-             patch("hindsight_api.main.create_app", return_value=mock_app), \
-             patch("hindsight_api.main.get_config") as mock_get_config, \
-             patch("hindsight_api.main.print_banner"), \
-             patch("uvicorn.run", side_effect=capture_uvicorn_run):
-
-            mock_config = MagicMock()
-            mock_config.host = "0.0.0.0"
-            mock_config.port = 8888
-            mock_config.log_level = "info"
-            mock_config.mcp_enabled = False
-            mock_config.run_migrations_on_startup = False
-            mock_config.database_url = "postgresql://test:test@localhost/test"
-            mock_get_config.return_value = mock_config
-            mock_engine.return_value = MagicMock()
-
-            with patch.object(sys, 'argv', ['hindsight-api', '--workers', '1']):
-                from hindsight_api.main import main
-                main()
-
-        assert len(uvicorn_calls) == 1
-        # With workers=1, should pass app object, not import string
-        assert uvicorn_calls[0]["app"] is mock_app, \
-            "main.py should pass app object (not import string) when workers=1"
-
-    def test_main_uses_import_string_for_multiple_workers(self, monkeypatch):
-        """
-        Verify that main.py uses import string when workers > 1.
-
-        This is important because multi-worker mode requires server.py to be imported
-        by each worker process.
-        """
-        monkeypatch.setenv("HINDSIGHT_API_WORKERS", "2")
-        monkeypatch.delenv("HINDSIGHT_API_TENANT_EXTENSION", raising=False)
-
-        uvicorn_calls = []
-
-        def capture_uvicorn_run(**kwargs):
-            uvicorn_calls.append(kwargs)
-
-        with patch("hindsight_api.main.MemoryEngine") as mock_engine, \
-             patch("hindsight_api.main.create_app") as mock_create_app, \
-             patch("hindsight_api.main.get_config") as mock_get_config, \
-             patch("hindsight_api.main.print_banner"), \
-             patch("uvicorn.run", side_effect=capture_uvicorn_run):
-
-            mock_config = MagicMock()
-            mock_config.host = "0.0.0.0"
-            mock_config.port = 8888
-            mock_config.log_level = "info"
-            mock_config.mcp_enabled = False
-            mock_config.run_migrations_on_startup = False
-            mock_config.database_url = "postgresql://test:test@localhost/test"
-            mock_get_config.return_value = mock_config
-            mock_engine.return_value = MagicMock()
-            mock_create_app.return_value = MagicMock()
-
-            with patch.object(sys, 'argv', ['hindsight-api', '--workers', '2']):
-                from hindsight_api.main import main
-                main()
-
-        assert len(uvicorn_calls) == 1
-        # With workers > 1, should use import string
-        assert uvicorn_calls[0]["app"] == "hindsight_api.server:app", \
-            "main.py should use import string when workers > 1"
-        assert uvicorn_calls[0]["workers"] == 2
-
-
-# Mock extensions for testing
-from hindsight_api.extensions import (
-    OperationValidatorExtension,
-    RecallContext,
-    ReflectContext,
-    RequestContext,
-    RetainContext,
-    TenantContext,
-    TenantExtension,
-    ValidationResult,
-)
-
-
-class MockTenantExtension(TenantExtension):
-    """Mock tenant extension for testing main.py extension loading."""
-
-    def __init__(self, config: dict):
-        super().__init__(config)
-        self._context_set = False
-
-    async def authenticate(self, request_context: RequestContext) -> TenantContext:
-        return TenantContext(schema_name="public")
-
-    async def list_tenants(self) -> list:
-        from hindsight_api.extensions.tenant import Tenant
-
-        return [Tenant(schema="public")]
-
-    def set_context(self, context) -> None:
-        self._context_set = True
-
-
-class MockOperationValidator(OperationValidatorExtension):
-    """Mock operation validator for testing main.py extension loading."""
-
-    def __init__(self, config: dict):
-        super().__init__(config)
-
-    async def validate_retain(self, ctx: RetainContext) -> ValidationResult:
-        return ValidationResult.accept()
-
-    async def validate_recall(self, ctx: RecallContext) -> ValidationResult:
-        return ValidationResult.accept()
-
-    async def validate_reflect(self, ctx: ReflectContext) -> ValidationResult:
-        return ValidationResult.accept()
@@ -62,9 +62,9 @@ async def test_local_mcp_server_recall(mock_memory):
    tools = mcp_server._tool_manager._tools
    assert "recall" in tools

-    # Call recall
+    # Call recall with new params
    recall_tool = tools["recall"]
-    result = await recall_tool.fn(query="test query", max_tokens=2048)
+    result = await recall_tool.fn(query="test query", max_tokens=2048, budget="mid")

    # Result is a dict
    assert isinstance(result, dict)
@@ -75,7 +75,7 @@ async def test_local_mcp_server_recall(mock_memory):
    assert call_kwargs["bank_id"] == "test-bank"
    assert call_kwargs["query"] == "test query"
    assert call_kwargs["max_tokens"] == 2048
-    assert call_kwargs["budget"] == Budget.HIGH
+    assert call_kwargs["budget"] == Budget.MID


@pytest.mark.asyncio
@@ -141,7 +141,7 @@ async def test_local_mcp_server_recall_error_handling(mock_memory):

@pytest.mark.asyncio
 async def test_local_mcp_server_recall_with_defaults(mock_memory):
-    """Test that recall uses default max_tokens and HIGH budget."""
+    """Test that recall uses default max_tokens and budget."""
    from hindsight_api.mcp_local import create_local_mcp_server
    from hindsight_api.engine.memory_engine import Budget

@@ -159,54 +159,4 @@ async def test_local_mcp_server_recall_with_defaults(mock_memory):

    call_kwargs = mock_memory.recall_async.call_args.kwargs
    assert call_kwargs["max_tokens"] == 4096
-    assert call_kwargs["budget"] == Budget.HIGH
-
-
-@pytest.mark.asyncio
-async def test_local_mcp_server_retain_with_timestamp(mock_memory):
-    """Test that retain passes timestamp as event_date."""
-    from datetime import datetime, timezone
-    from hindsight_api.mcp_local import create_local_mcp_server
-
-    mcp_server = create_local_mcp_server("test-bank", memory=mock_memory)
-
-    tools = mcp_server._tool_manager._tools
-    retain_tool = tools["retain"]
-
-    # Call retain with timestamp
-    result = await retain_tool.fn(
-        content="test content", context="test_context", timestamp="2024-01-15T10:30:00Z"
-    )
-
-    assert result["status"] == "accepted"
-
-    # Wait for background task
-    await asyncio.sleep(0.1)
-
-    call_kwargs = mock_memory.retain_batch_async.call_args.kwargs
-    contents = call_kwargs["contents"]
-    assert len(contents) == 1
-    assert contents[0]["content"] == "test content"
-    assert contents[0]["context"] == "test_context"
-    assert "event_date" in contents[0]
-    assert contents[0]["event_date"] == datetime(2024, 1, 15, 10, 30, 0, tzinfo=timezone.utc)
-
-
-@pytest.mark.asyncio
-async def test_local_mcp_server_retain_with_invalid_timestamp(mock_memory):
-    """Test that retain rejects invalid timestamp format."""
-    from hindsight_api.mcp_local import create_local_mcp_server
-
-    mcp_server = create_local_mcp_server("test-bank", memory=mock_memory)
-
-    tools = mcp_server._tool_manager._tools
-    retain_tool = tools["retain"]
-
-    # Call retain with invalid timestamp
-    result = await retain_tool.fn(content="test content", timestamp="not-a-date")
-
-    assert result["status"] == "error"
-    assert "Invalid timestamp format" in result["message"]
-
-    # Verify retain_batch_async was NOT called
-    mock_memory.retain_batch_async.assert_not_called()
+    assert call_kwargs["budget"] == Budget.LOW
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Nicolò Boschi	393add896e	fix: batch queries on recall	2026-01-13 11:54:54 +01:00
Nicolò Boschi	f1dccda218	fix: batch queries on recall	2026-01-13 11:40:48 +01:00