doc: show embed page

Release v0.4.5
- Update version to 0.4.5 in all components - Python packages: hindsight-api, hindsight-dev, hindsight-all, hindsight-litellm, hindsight-embed - Python client: hindsight-clients/python - TypeScript client: hindsight-clients/typescript - Rust CLI: hindsight-cli - Control Plane: hindsight-control-plane - OpenClaw integration: hindsight-integrations/openclaw - Helm chart - Sync documentation to version-0.4
2026-01-30 15:53:17 +01:00 · 2026-01-30 14:55:50 +01:00 · 2026-01-30 14:54:32 +01:00 · 2026-01-30 13:05:07 +01:00 · 2026-01-30 13:04:42 +01:00 · 2026-01-30 13:04:33 +01:00
149 changed files with 1942 additions and 1188 deletions
@@ -139,7 +139,7 @@ jobs:
        path: hindsight-clients/typescript/*.tgz
        retention-days: 1

-  release-moltbot-integration:
+  release-openclaw-integration:
    runs-on: ubuntu-latest
    environment: npm

@@ -153,15 +153,15 @@ jobs:
        registry-url: 'https://registry.npmjs.org'

    - name: Install dependencies
-      working-directory: ./hindsight-integrations/moltbot
+      working-directory: ./hindsight-integrations/openclaw
      run: npm ci

    - name: Build
-      working-directory: ./hindsight-integrations/moltbot
+      working-directory: ./hindsight-integrations/openclaw
      run: npm run build

    - name: Publish to npm
-      working-directory: ./hindsight-integrations/moltbot
+      working-directory: ./hindsight-integrations/openclaw
      run: |
        set +e
        OUTPUT=$(npm publish --access public 2>&1)
@@ -178,14 +178,14 @@ jobs:
        NODE_AUTH_TOKEN: ${{ secrets.NPM_TOKEN }}

    - name: Pack for GitHub release
-      working-directory: ./hindsight-integrations/moltbot
+      working-directory: ./hindsight-integrations/openclaw
      run: npm pack

    - name: Upload artifacts
      uses: actions/upload-artifact@v4
      with:
-        name: moltbot-integration
-        path: hindsight-integrations/moltbot/*.tgz
+        name: openclaw-integration
+        path: hindsight-integrations/openclaw/*.tgz
        retention-days: 1

  release-control-plane:
@@ -415,7 +415,7 @@ jobs:

  create-github-release:
    runs-on: ubuntu-latest
-    needs: [release-python-packages, release-typescript-client, release-moltbot-integration, release-control-plane, release-rust-cli, release-docker-images, release-helm-chart]
+    needs: [release-python-packages, release-typescript-client, release-openclaw-integration, release-control-plane, release-rust-cli, release-docker-images, release-helm-chart]
    permissions:
      contents: write

@@ -438,11 +438,11 @@ jobs:
        name: typescript-client
        path: ./artifacts/typescript-client

-    - name: Download Moltbot Integration
+    - name: Download OpenClaw Integration
      uses: actions/download-artifact@v4
      with:
-        name: moltbot-integration
-        path: ./artifacts/moltbot-integration
+        name: openclaw-integration
+        path: ./artifacts/openclaw-integration

    - name: Download Control Plane
      uses: actions/download-artifact@v4
@@ -485,8 +485,8 @@ jobs:
        cp artifacts/python-packages/hindsight-embed/dist/* release-assets/ || true
        # TypeScript client
        cp artifacts/typescript-client/*.tgz release-assets/ || true
-        # Moltbot Integration
-        cp artifacts/moltbot-integration/*.tgz release-assets/ || true
+        # OpenClaw Integration
+        cp artifacts/openclaw-integration/*.tgz release-assets/ || true
        # Control Plane
        cp artifacts/control-plane/*.tgz release-assets/ || true
        # Rust CLI binaries
@@ -82,7 +82,7 @@ jobs:
    - name: Build TypeScript client
      run: npm run build --workspace=hindsight-clients/typescript

-  build-moltbot-integration:
+  build-openclaw-integration:
    runs-on: ubuntu-latest

    steps:
@@ -94,15 +94,15 @@ jobs:
        node-version: '22'

    - name: Install dependencies
-      working-directory: ./hindsight-integrations/moltbot
+      working-directory: ./hindsight-integrations/openclaw
      run: npm ci

    - name: Run tests
-      working-directory: ./hindsight-integrations/moltbot
+      working-directory: ./hindsight-integrations/openclaw
      run: npm test

    - name: Build
-      working-directory: ./hindsight-integrations/moltbot
+      working-directory: ./hindsight-integrations/openclaw
      run: npm run build

  build-control-plane:
@@ -93,6 +93,34 @@ uv run ty check hindsight_api  # Type check
 3. Run tests to ensure nothing breaks
 4. Submit a PR with a clear description of changes

+## Release Process
+
+The project uses `scripts/release.sh` for creating releases. This script automates the entire release workflow:
+
+1. Bumps version in all components (API, clients, CLI, control plane, Helm)
+2. **Regenerates OpenAPI spec and client SDKs** (Python, TypeScript, Rust)
+3. Updates documentation versioning
+4. Creates a commit and git tag
+5. Pushes to GitHub (triggers CI/CD to publish packages)
+
+### Usage
+
+```bash
+./scripts/release.sh <version>
+```
+
+**Example:**
+```bash
+./scripts/release.sh 0.5.0
+```
+
+### Important for Developers
+
+- During development, version bumps in `__init__.py` do NOT require client regeneration
+- Clients are only regenerated during releases
+- Do not manually run `./scripts/generate-clients.sh` unless testing generation changes
+- Client version comments will reflect the API version from the latest release
+
 ## Reporting Issues

 Open an issue on GitHub with:
@@ -169,16 +169,33 @@ ENV PATH="/app/api/.venv/bin:${PATH}"

 # Pre-download ML models to avoid runtime download (conditional)
 # Only runs if both PRELOAD_ML_MODELS=true AND INCLUDE_LOCAL_MODELS=true
+# Includes retry logic with exponential backoff for transient network failures
 ARG PRELOAD_ML_MODELS
 ARG INCLUDE_LOCAL_MODELS
+ENV HF_HUB_DOWNLOAD_TIMEOUT=600
 RUN if [ "$PRELOAD_ML_MODELS" = "true" ] && [ "$INCLUDE_LOCAL_MODELS" = "true" ]; then \
-    /app/api/.venv/bin/python -c "\
+    MAX_RETRIES=3; \
+    RETRY_DELAY=10; \
+    for i in $(seq 1 $MAX_RETRIES); do \
+      echo "Attempt $i/$MAX_RETRIES: Downloading ML models..."; \
+      /app/api/.venv/bin/python -c "\
+import os; os.environ['HF_HUB_DOWNLOAD_TIMEOUT'] = '600'; \
 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 cached successfully')"; \
+print('Models cached successfully')" && break; \
+      if [ $i -lt $MAX_RETRIES ]; then \
+        echo "Attempt $i failed, retrying in ${RETRY_DELAY}s..."; \
+        sleep $RETRY_DELAY; \
+        RETRY_DELAY=$((RETRY_DELAY * 2)); \
+      fi; \
+    done; \
+    if [ $i -eq $MAX_RETRIES ] && ! /app/api/.venv/bin/python -c "from sentence_transformers import SentenceTransformer; SentenceTransformer('BAAI/bge-small-en-v1.5')" 2>/dev/null; then \
+      echo "ERROR: Failed to download models after $MAX_RETRIES attempts"; \
+      exit 1; \
+    fi; \
    elif [ "$INCLUDE_LOCAL_MODELS" != "true" ]; then echo "Skipping ML model preload (local-models not included)"; \
    else echo "Skipping ML model preload"; fi

@@ -277,16 +294,33 @@ ENV PATH="/app/api/.venv/bin:${PATH}"

 # Pre-download ML models to avoid runtime download (conditional)
 # Only runs if both PRELOAD_ML_MODELS=true AND INCLUDE_LOCAL_MODELS=true
+# Includes retry logic with exponential backoff for transient network failures
 ARG PRELOAD_ML_MODELS
 ARG INCLUDE_LOCAL_MODELS
+ENV HF_HUB_DOWNLOAD_TIMEOUT=600
 RUN if [ "$PRELOAD_ML_MODELS" = "true" ] && [ "$INCLUDE_LOCAL_MODELS" = "true" ]; then \
-    /app/api/.venv/bin/python -c "\
+    MAX_RETRIES=3; \
+    RETRY_DELAY=10; \
+    for i in $(seq 1 $MAX_RETRIES); do \
+      echo "Attempt $i/$MAX_RETRIES: Downloading ML models..."; \
+      /app/api/.venv/bin/python -c "\
+import os; os.environ['HF_HUB_DOWNLOAD_TIMEOUT'] = '600'; \
 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 cached successfully')"; \
+print('Models cached successfully')" && break; \
+      if [ $i -lt $MAX_RETRIES ]; then \
+        echo "Attempt $i failed, retrying in ${RETRY_DELAY}s..."; \
+        sleep $RETRY_DELAY; \
+        RETRY_DELAY=$((RETRY_DELAY * 2)); \
+      fi; \
+    done; \
+    if [ $i -eq $MAX_RETRIES ] && ! /app/api/.venv/bin/python -c "from sentence_transformers import SentenceTransformer; SentenceTransformer('BAAI/bge-small-en-v1.5')" 2>/dev/null; then \
+      echo "ERROR: Failed to download models after $MAX_RETRIES attempts"; \
+      exit 1; \
+    fi; \
    elif [ "$INCLUDE_LOCAL_MODELS" != "true" ]; then echo "Skipping ML model preload (local-models not included)"; \
    else echo "Skipping ML model preload"; fi

@@ -2,8 +2,8 @@ apiVersion: v2
 name: hindsight
 description: Hindsight helm chart
 type: application
-version: 0.4.2
-appVersion: "0.4.2"
+version: 0.4.5
+appVersion: "0.4.5"
 keywords:
  - ai
  - memory
@@ -46,4 +46,4 @@ __all__ = [
    "RemoteTEICrossEncoder",
    "LLMConfig",
 ]
-__version__ = "0.4.2"
+__version__ = "0.4.5"
@@ -1404,9 +1404,10 @@ def create_app(
                worker_id=worker_id,
                executor=memory.execute_task,
                poll_interval_ms=config.worker_poll_interval_ms,
-                batch_size=config.worker_batch_size,
                max_retries=config.worker_max_retries,
                tenant_extension=getattr(memory, "_tenant_extension", None),
+                max_slots=config.worker_max_slots,
+                consolidation_max_slots=config.worker_consolidation_max_slots,
            )
            poller_task = asyncio.create_task(poller.run())
            logging.info(f"Worker poller started (worker_id={worker_id})")
@@ -143,8 +143,9 @@ 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"
+ENV_WORKER_MAX_SLOTS = "HINDSIGHT_API_WORKER_MAX_SLOTS"
+ENV_WORKER_CONSOLIDATION_MAX_SLOTS = "HINDSIGHT_API_WORKER_CONSOLIDATION_MAX_SLOTS"

 # Reflect agent settings
 ENV_REFLECT_MAX_ITERATIONS = "HINDSIGHT_API_REFLECT_MAX_ITERATIONS"
@@ -229,8 +230,9 @@ 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
+DEFAULT_WORKER_MAX_SLOTS = 10  # Total concurrent tasks per worker
+DEFAULT_WORKER_CONSOLIDATION_MAX_SLOTS = 2  # Max concurrent consolidation tasks per worker

 # Reflect agent settings
 DEFAULT_REFLECT_MAX_ITERATIONS = 10  # Max tool call iterations before forcing response
@@ -419,8 +421,9 @@ class HindsightConfig:
    worker_id: str | None
    worker_poll_interval_ms: int
    worker_max_retries: int
-    worker_batch_size: int
    worker_http_port: int
+    worker_max_slots: int
+    worker_consolidation_max_slots: int

    # Reflect agent settings
    reflect_max_iterations: int
@@ -582,8 +585,11 @@ class HindsightConfig:
            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))),
+            worker_max_slots=int(os.getenv(ENV_WORKER_MAX_SLOTS, str(DEFAULT_WORKER_MAX_SLOTS))),
+            worker_consolidation_max_slots=int(
+                os.getenv(ENV_WORKER_CONSOLIDATION_MAX_SLOTS, str(DEFAULT_WORKER_CONSOLIDATION_MAX_SLOTS))
+            ),
            # Reflect agent settings
            reflect_max_iterations=int(os.getenv(ENV_REFLECT_MAX_ITERATIONS, str(DEFAULT_REFLECT_MAX_ITERATIONS))),
        )
@@ -759,6 +759,10 @@ async def _extract_facts_from_chunk(

    # Build user message with metadata and chunk content in a clear format
    # Format event_date with day of week for better temporal reasoning
+    # Handle both datetime objects and ISO string formats (from deserialized async tasks)
+    from .orchestrator import parse_datetime_flexible
+
+    event_date = parse_datetime_flexible(event_date)
    event_date_formatted = event_date.strftime("%A, %B %d, %Y")  # e.g., "Monday, June 10, 2024"
    user_message = f"""Extract facts from the following text chunk.

@@ -1346,6 +1350,8 @@ def _add_temporal_offsets(facts: list[ExtractedFactType], contents: list[RetainC

    Modifies facts in place.
    """
+    from .orchestrator import parse_datetime_flexible
+
    # Group facts by content_index
    current_content_idx = 0
    content_fact_start = 0
@@ -1360,10 +1366,10 @@ def _add_temporal_offsets(facts: list[ExtractedFactType], contents: list[RetainC
        fact_position = i - content_fact_start
        offset = timedelta(seconds=fact_position * SECONDS_PER_FACT)

-        # Apply offset to all temporal fields
+        # Apply offset to all temporal fields (handle both datetime objects and ISO strings)
        if fact.occurred_start:
-            fact.occurred_start = fact.occurred_start + offset
+            fact.occurred_start = parse_datetime_flexible(fact.occurred_start) + offset
        if fact.occurred_end:
-            fact.occurred_end = fact.occurred_end + offset
+            fact.occurred_end = parse_datetime_flexible(fact.occurred_end) + offset
        if fact.mentioned_at:
-            fact.mentioned_at = fact.mentioned_at + offset
+            fact.mentioned_at = parse_datetime_flexible(fact.mentioned_at) + offset
@@ -8,6 +8,7 @@ import logging
 import time
 import uuid
 from datetime import UTC, datetime
+from typing import Any

 from ..db_utils import acquire_with_retry
 from . import bank_utils
@@ -18,6 +19,39 @@ def utcnow():
    return datetime.now(UTC)


+def parse_datetime_flexible(value: Any) -> datetime:
+    """
+    Parse a datetime value that could be either a datetime object or an ISO string.
+
+    This handles datetime values from both direct Python calls and deserialized JSON
+    (where datetime objects are serialized as ISO strings).
+
+    Args:
+        value: Either a datetime object or an ISO format string
+
+    Returns:
+        datetime object (timezone-aware)
+
+    Raises:
+        TypeError: If value is neither datetime nor string
+        ValueError: If string is not a valid ISO datetime
+    """
+    if isinstance(value, datetime):
+        # Ensure timezone-aware
+        if value.tzinfo is None:
+            return value.replace(tzinfo=UTC)
+        return value
+    elif isinstance(value, str):
+        # Parse ISO format string (handles both 'Z' and '+00:00' timezone formats)
+        dt = datetime.fromisoformat(value.replace("Z", "+00:00"))
+        # Ensure timezone-aware
+        if dt.tzinfo is None:
+            return dt.replace(tzinfo=UTC)
+        return dt
+    else:
+        raise TypeError(f"Expected datetime or string, got {type(value).__name__}")
+
+
 from ..response_models import TokenUsage
 from . import (
    chunk_storage,
@@ -89,10 +123,18 @@ async def retain_batch(
        # Merge item-level tags with document-level tags
        item_tags = item.get("tags", []) or []
        merged_tags = list(set(item_tags + (document_tags or [])))
+
+        # Handle event_date: parse flexibly (handles both datetime objects and ISO strings)
+        event_date_value = item.get("event_date")
+        if event_date_value:
+            event_date_value = parse_datetime_flexible(event_date_value)
+        else:
+            event_date_value = utcnow()
+
        content = RetainContent(
            content=item["content"],
            context=item.get("context", ""),
-            event_date=item.get("event_date") or utcnow(),
+            event_date=event_date_value,
            metadata=item.get("metadata", {}),
            entities=item.get("entities", []),
            tags=merged_tags,
@@ -182,7 +182,16 @@ class BrokerTaskBackend(TaskBackend):
        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)
+
+        # Custom encoder to handle datetime objects
+        from datetime import datetime
+
+        def datetime_encoder(obj):
+            if isinstance(obj, datetime):
+                return obj.isoformat()
+            raise TypeError(f"Object of type {type(obj).__name__} is not JSON serializable")
+
+        payload_json = json.dumps(task_dict, default=datetime_encoder)

        schema = self._schema_getter() if self._schema_getter else self._schema
        table = fq_table("async_operations", schema)
@@ -253,8 +253,9 @@ def main():
            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,
+            worker_max_slots=config.worker_max_slots,
+            worker_consolidation_max_slots=config.worker_consolidation_max_slots,
            reflect_max_iterations=config.reflect_max_iterations,
            mental_model_refresh_concurrency=config.mental_model_refresh_concurrency,
        )
@@ -124,12 +124,6 @@ def main():
        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,
@@ -168,8 +162,9 @@ def main():

    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"  Max slots: {config.worker_max_slots}")
+    print(f"  Consolidation max slots: {config.worker_consolidation_max_slots}")
    print(f"  HTTP server: {args.http_host}:{args.http_port}")
    print()

@@ -213,9 +208,10 @@ def main():
            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,
+            max_slots=config.worker_max_slots,
+            consolidation_max_slots=config.worker_consolidation_max_slots,
        )

        # Create the HTTP app for metrics/health
@@ -57,10 +57,11 @@ class WorkerPoller:
        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,
+        max_slots: int = 10,
+        consolidation_max_slots: int = 2,
    ):
        """
        Initialize the worker poller.
@@ -70,28 +71,32 @@ class WorkerPoller:
            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.
+            max_slots: Maximum concurrent tasks per worker
+            consolidation_max_slots: Maximum concurrent consolidation tasks per worker
        """
        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._max_slots = max_slots
+        self._consolidation_max_slots = consolidation_max_slots
        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]] = {}
+        # Track active tasks locally: operation_id -> (op_type, bank_id, schema, asyncio.Task)
+        self._active_tasks: dict[str, tuple[str, str, str | None, asyncio.Task]] = {}
+        # Track in-flight tasks by operation type
+        self._in_flight_by_type: dict[str, int] = {}

    async def _get_schemas(self) -> list[str | None]:
        """Get list of schemas to poll. Returns [None] for public schema."""
@@ -102,67 +107,114 @@ class WorkerPoller:
        # Single schema mode
        return [self._schema]

+    async def _get_available_slots(self) -> tuple[int, int]:
+        """
+        Calculate available slots for claiming tasks.
+
+        Returns:
+            (total_available, consolidation_available) tuple
+        """
+        async with self._in_flight_lock:
+            total_in_flight = self._in_flight_count
+            consolidation_in_flight = self._in_flight_by_type.get("consolidation", 0)
+
+        total_available = max(0, self._max_slots - total_in_flight)
+        consolidation_available = max(0, self._consolidation_max_slots - consolidation_in_flight)
+
+        return total_available, consolidation_available
+
+    async def wait_for_active_tasks(self, timeout: float = 10.0) -> bool:
+        """
+        Wait for all active background tasks to complete (test helper).
+
+        This is a test-only utility that allows tests to synchronize with
+        fire-and-forget background tasks without using sleep().
+
+        Args:
+            timeout: Maximum time to wait in seconds
+
+        Returns:
+            True if all tasks completed, False if timeout was reached
+        """
+        start_time = asyncio.get_event_loop().time()
+        while True:
+            async with self._in_flight_lock:
+                if self._in_flight_count == 0:
+                    return True
+
+            elapsed = asyncio.get_event_loop().time() - start_time
+            if elapsed >= timeout:
+                return False
+
+            # Short sleep to avoid busy-waiting
+            await asyncio.sleep(0.01)
+
    async def claim_batch(self) -> list[ClaimedTask]:
        """
-        Claim up to batch_size pending tasks atomically across all tenant schemas.
+        Claim pending tasks atomically across all tenant schemas,
+        respecting slot limits (total and consolidation).

        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
        """
+        # Calculate available slots
+        total_available, consolidation_available = await self._get_available_slots()
+
+        if total_available <= 0:
+            return []
+
        schemas = await self._get_schemas()
        all_tasks: list[ClaimedTask] = []
-        remaining_batch = self._batch_size
+        remaining_total = total_available
+        remaining_consolidation = consolidation_available

        for schema in schemas:
-            if remaining_batch <= 0:
+            if remaining_total <= 0:
                break

-            tasks = await self._claim_batch_for_schema(schema, remaining_batch)
+            tasks = await self._claim_batch_for_schema(schema, remaining_total, remaining_consolidation)
+
+            # Update remaining slots based on what was claimed
+            for task in tasks:
+                op_type = task.task_dict.get("operation_type", "unknown")
+                if op_type == "consolidation":
+                    remaining_consolidation -= 1
+
            all_tasks.extend(tasks)
-            remaining_batch -= len(tasks)
+            remaining_total -= 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."""
+    async def _claim_batch_for_schema(
+        self, schema: str | None, limit: int, consolidation_limit: int
+    ) -> list[ClaimedTask]:
+        """Claim tasks from a specific schema respecting slot limits."""
        try:
-            return await self._claim_batch_for_schema_inner(schema, limit)
+            return await self._claim_batch_for_schema_inner(schema, limit, consolidation_limit)
        except Exception as e:
            logger.warning(f"Worker {self._worker_id} failed to claim tasks for schema {schema or 'public'}: {e}")
            return []

-    async def _claim_batch_for_schema_inner(self, schema: str | None, limit: int) -> list[ClaimedTask]:
-        """Inner implementation for claiming tasks from a specific schema."""
+    async def _claim_batch_for_schema_inner(
+        self, schema: str | None, limit: int, consolidation_limit: int
+    ) -> list[ClaimedTask]:
+        """Inner implementation for claiming tasks from a specific schema with slot limits."""
        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(
+                # Strategy: Claim non-consolidation tasks first, then consolidation up to limit
+
+                # 1. Claim non-consolidation tasks (up to limit)
+                non_consolidation_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'
-                        )
-                    )
+                    FROM {table}
+                    WHERE status = 'pending'
+                      AND task_payload IS NOT NULL
+                      AND operation_type != 'consolidation'
                    ORDER BY created_at
                    LIMIT $1
                    FOR UPDATE SKIP LOCKED
@@ -170,11 +222,39 @@ class WorkerPoller:
                    limit,
                )

-                if not rows:
+                claimed_count = len(non_consolidation_rows)
+                remaining_limit = limit - claimed_count
+
+                # 2. Claim consolidation tasks (up to consolidation_limit and remaining_limit)
+                consolidation_rows = []
+                if consolidation_limit > 0 and remaining_limit > 0:
+                    consolidation_rows = await conn.fetch(
+                        f"""
+                        SELECT operation_id, task_payload
+                        FROM {table} AS pending
+                        WHERE status = 'pending'
+                          AND task_payload IS NOT NULL
+                          AND operation_type = 'consolidation'
+                          AND 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
+                        """,
+                        min(consolidation_limit, remaining_limit),
+                    )
+
+                all_rows = non_consolidation_rows + consolidation_rows
+
+                if not all_rows:
                    return []

                # Claim the tasks by updating status and worker_id
-                operation_ids = [row["operation_id"] for row in rows]
+                operation_ids = [row["operation_id"] for row in all_rows]
                await conn.execute(
                    f"""
                    UPDATE {table}
@@ -192,7 +272,7 @@ class WorkerPoller:
                        task_dict=json.loads(row["task_payload"]),
                        schema=schema,
                    )
-                    for row in rows
+                    for row in all_rows
                ]

    async def _mark_completed(self, operation_id: str, schema: str | None):
@@ -258,18 +338,43 @@ class WorkerPoller:
            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."""
+        """Execute a single task as a background job (fire-and-forget)."""
        task_type = task.task_dict.get("type", "unknown")
+        operation_type = task.task_dict.get("operation_type", "unknown")
        bank_id = task.task_dict.get("bank_id", "unknown")

+        # Create background task
+        bg_task = asyncio.create_task(self._execute_task_inner(task))
+
        # Track this task as active
        async with self._in_flight_lock:
-            self._active_tasks[task.operation_id] = (task_type, bank_id, task.schema)
+            self._active_tasks[task.operation_id] = (task_type, bank_id, task.schema, bg_task)
+            self._in_flight_count += 1
+            self._in_flight_by_type[operation_type] = self._in_flight_by_type.get(operation_type, 0) + 1
+
+        # Add cleanup callback
+        bg_task.add_done_callback(lambda _: asyncio.create_task(self._cleanup_task(task.operation_id, operation_type)))
+
+    async def _cleanup_task(self, operation_id: str, operation_type: str):
+        """Remove task from tracking after completion."""
+        async with self._in_flight_lock:
+            if operation_id in self._active_tasks:
+                self._active_tasks.pop(operation_id, None)
+                self._in_flight_count -= 1
+                count = self._in_flight_by_type.get(operation_type, 0)
+                if count > 0:
+                    self._in_flight_by_type[operation_type] = count - 1
+                    if self._in_flight_by_type[operation_type] == 0:
+                        del self._in_flight_by_type[operation_type]
+
+    async def _execute_task_inner(self, task: ClaimedTask):
+        """Inner task execution with error handling."""
+        task_type = task.task_dict.get("type", "unknown")
+        bank_id = task.task_dict.get("bank_id", "unknown")

        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)
@@ -279,10 +384,6 @@ class WorkerPoller:
            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:
        """
@@ -325,59 +426,59 @@ class WorkerPoller:

    async def run(self):
        """
-        Main polling loop.
+        Main polling loop with fire-and-forget task execution.

-        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.
+        Continuously polls for pending tasks, spawns them as background tasks,
+        and immediately continues polling (up to slot limits).
        """
-        # Recover any tasks from a previous crash before starting
        await self.recover_own_tasks()

-        logger.info(f"Worker {self._worker_id} starting polling loop")
+        logger.info(
+            f"Worker {self._worker_id} starting polling loop "
+            f"(max_slots={self._max_slots}, consolidation_max_slots={self._consolidation_max_slots})"
+        )

        while not self._shutdown.is_set():
            try:
-                # Claim a batch of tasks (across all tenant schemas if configured)
+                # Claim a batch of tasks (respecting slot limits)
                tasks = await self.claim_batch()

                if tasks:
                    # Log batch info
                    task_types: dict[str, int] = {}
                    schemas_seen: set[str | None] = set()
+                    consolidation_count = 0
                    for task in tasks:
                        t = task.task_dict.get("type", "unknown")
+                        op_type = task.task_dict.get("operation_type", "unknown")
                        task_types[t] = task_types.get(t, 0) + 1
                        schemas_seen.add(task.schema)
+                        if op_type == "consolidation":
+                            consolidation_count += 1
+
                    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})"
+                        f"Worker {self._worker_id} claimed {len(tasks)} tasks "
+                        f"({consolidation_count} consolidation): {types_str} (schemas: {schemas_str})"
                    )

-                    # Track in-flight tasks
-                    async with self._in_flight_lock:
-                        self._in_flight_count += len(tasks)
+                    # Spawn tasks as background jobs (fire-and-forget)
+                    for task in tasks:
+                        await self.execute_task(task)

-                    # 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
+                    # Continue immediately to claim more tasks (if slots available)
+                    continue
+
+                # No tasks claimed (either no pending tasks or slots full)
+                # 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()
@@ -408,15 +509,27 @@ class WorkerPoller:
        while asyncio.get_event_loop().time() - start_time < timeout:
            async with self._in_flight_lock:
                in_flight = self._in_flight_count
+                active_task_objects = [task_info[3] for task_info in self._active_tasks.values()]

            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")
+            # Wait for at least one task to complete
+            if active_task_objects:
+                done, _ = await asyncio.wait(active_task_objects, timeout=0.5, return_when=asyncio.FIRST_COMPLETED)
+            else:
+                await asyncio.sleep(0.5)
+
+        logger.warning(f"Worker {self._worker_id} shutdown timeout after {timeout}s, cancelling remaining tasks")
+
+        # Cancel remaining tasks
+        async with self._in_flight_lock:
+            for operation_id, (_, _, _, bg_task) in list(self._active_tasks.items()):
+                if not bg_task.done():
+                    bg_task.cancel()

    async def _log_progress_if_due(self):
        """Log progress stats every PROGRESS_LOG_INTERVAL seconds."""
@@ -427,14 +540,19 @@ class WorkerPoller:
        self._last_progress_log = now

        try:
-            # Get local active tasks (this worker only)
+            # Get local active tasks
            async with self._in_flight_lock:
                in_flight = self._in_flight_count
-                active_tasks = dict(self._active_tasks)  # Copy to avoid holding lock
+                in_flight_by_type = dict(self._in_flight_by_type)
+                active_tasks = dict(self._active_tasks)

-            # Build local processing breakdown grouped by (op_type, bank_id)
+            consolidation_count = in_flight_by_type.get("consolidation", 0)
+            available_slots = self._max_slots - in_flight
+            available_consolidation_slots = self._consolidation_max_slots - consolidation_count
+
+            # Build local processing breakdown
            task_groups: dict[tuple[str, str], int] = {}
-            for op_type, bank_id, _ in active_tasks.values():
+            for op_type, bank_id, _, _ in active_tasks.values():
                key = (op_type, bank_id)
                task_groups[key] = task_groups.get(key, 0) + 1

@@ -443,7 +561,7 @@ class WorkerPoller:
            if len(processing_info) > 10:
                processing_str += f" +{len(processing_info) - 10} more"

-            # Get global stats from DB across all schemas
+            # Get global stats from DB
            schemas = await self._get_schemas()
            global_pending = 0
            all_worker_counts: dict[str, int] = {}
@@ -455,7 +573,6 @@ class WorkerPoller:
                    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
@@ -468,7 +585,6 @@ class WorkerPoller:
                        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:
@@ -477,7 +593,9 @@ class WorkerPoller:

            schemas_str = ", ".join(s or "public" for s in schemas)
            logger.info(
-                f"[WORKER_STATS] worker={self._worker_id} in_flight={in_flight} | "
+                f"[WORKER_STATS] worker={self._worker_id} "
+                f"slots={in_flight}/{self._max_slots} (consolidation={consolidation_count}/{self._consolidation_max_slots}) | "
+                f"available={available_slots} (consolidation={available_consolidation_slots}) | "
                f"global: pending={global_pending} (schemas: {schemas_str}) | "
                f"others: {others_str} | "
                f"my_active: {processing_str}"
@@ -4,7 +4,7 @@ build-backend = "hatchling.build"

 [project]
 name = "hindsight-api"
-version = "0.4.2"
+version = "0.4.5"
 description = "Hindsight: Agent Memory That Works Like Human Memory"
 readme = "README.md"
 requires-python = ">=3.11"
@@ -346,11 +346,11 @@ class TestConsolidationIntegration:
        or when one directly updates another (e.g., location change).

        Given:
-        - "Nicolò lives in Italy"
-        - "Nicolò moved to the US recently" (updates the living location)
+        - "Alex lives in Italy"
+        - "Alex moved to the US recently" (updates the living location)

        The second fact should UPDATE the first, not create a separate observation.
-        But unrelated facts like "Nicolò works at Vectorize" should stay separate.
+        But unrelated facts like "Alex works at Vectorize" should stay separate.
        """
        bank_id = f"test-consolidation-merge-{uuid.uuid4().hex[:8]}"

@@ -360,14 +360,14 @@ class TestConsolidationIntegration:
        # Retain a memory about living location
        await memory.retain_async(
            bank_id=bank_id,
-            content="Nicolò lives in Italy.",
+            content="Alex lives in Italy.",
            request_context=request_context,
        )

        # Retain an unrelated memory (different topic - should NOT merge)
        await memory.retain_async(
            bank_id=bank_id,
-            content="Nicolò works at Vectorize as an engineer.",
+            content="Alex works at Vectorize as an engineer.",
            request_context=request_context,
        )

@@ -384,7 +384,7 @@ class TestConsolidationIntegration:
        # Add a memory that UPDATES the living location (should merge with first)
        await memory.retain_async(
            bank_id=bank_id,
-            content="Nicolò recently moved to the United States.",
+            content="Alex recently moved to the United States.",
            request_context=request_context,
        )

@@ -485,9 +485,9 @@ class TestConsolidationIntegration:
        they should be merged into ONE observation that captures the change.

        Example:
-        - "Nicolò loves pizza"
-        - "Nicolò hates pizza"
-        → Should become: "Nicolò used to love pizza but now hates it" (or similar)
+        - "Alex loves pizza"
+        - "Alex hates pizza"
+        → Should become: "Alex used to love pizza but now hates it" (or similar)
        """
        bank_id = f"test-consolidation-contradict-{uuid.uuid4().hex[:8]}"

@@ -497,7 +497,7 @@ class TestConsolidationIntegration:
        # Add initial fact
        await memory.retain_async(
            bank_id=bank_id,
-            content="Nicolò loves pizza.",
+            content="Alex loves pizza.",
            request_context=request_context,
        )

@@ -515,7 +515,7 @@ class TestConsolidationIntegration:
        # Add contradicting fact (same person, same topic, opposite sentiment)
        await memory.retain_async(
            bank_id=bank_id,
-            content="Nicolò hates pizza.",
+            content="Alex hates pizza.",
            request_context=request_context,
        )

@@ -1098,3 +1098,134 @@ async def test_version_endpoint_returns_correct_version(api_client):
    assert isinstance(features["worker"], bool)

    print(f"Version endpoint returned: api_version={result['api_version']}, features={features}")
+
+
+@pytest.mark.asyncio
+async def test_retain_with_timestamp_async(api_client, test_bank_id):
+    """Test that async retain accepts timestamp field and serializes correctly."""
+    response = await api_client.post(
+        f"/v1/default/banks/{test_bank_id}/memories",
+        json={
+            "items": [
+                {
+                    "content": "Test memory with timestamp",
+                    "context": "test",
+                    "timestamp": "2026-01-30T11:45:00Z"
+                }
+            ],
+            "async": True
+        }
+    )
+    
+    assert response.status_code == 200, f"Expected 200, got {response.status_code}: {response.text}"
+    data = response.json()
+    assert data["success"] is True
+    assert data["async"] is True
+    assert "operation_id" in data
+    
+
+@pytest.mark.asyncio
+async def test_retain_with_timestamp_sync(api_client, test_bank_id):
+    """Test that sync retain accepts timestamp field."""
+    response = await api_client.post(
+        f"/v1/default/banks/{test_bank_id}/memories",
+        json={
+            "items": [
+                {
+                    "content": "Test memory with timestamp sync",
+                    "context": "test", 
+                    "timestamp": "2026-01-30T11:45:00Z"
+                }
+            ],
+            "async": False
+        }
+    )
+    
+    assert response.status_code == 200, f"Expected 200, got {response.status_code}: {response.text}"
+    data = response.json()
+    assert data["success"] is True
+    assert data["async"] is False
+    
+
+@pytest.mark.asyncio
+async def test_retain_with_multiple_timestamps(api_client, test_bank_id):
+    """Test that multiple items with different timestamp formats work."""
+    response = await api_client.post(
+        f"/v1/default/banks/{test_bank_id}/memories",
+        json={
+            "items": [
+                {
+                    "content": "Event 1",
+                    "timestamp": "2026-01-30T11:45:00Z"  # With Z
+                },
+                {
+                    "content": "Event 2", 
+                    "timestamp": "2026-01-30T12:00:00+00:00"  # With timezone
+                },
+                {
+                    "content": "Event 3"  # No timestamp
+                }
+            ],
+            "async": True
+        }
+    )
+    
+    assert response.status_code == 200, f"Expected 200, got {response.status_code}: {response.text}"
+    data = response.json()
+    assert data["success"] is True
+    assert data["items_count"] == 3
+
+
+@pytest.mark.asyncio
+async def test_retain_with_timestamp_async_complete_processing(api_client, test_bank_id):
+    """Test that async retain with timestamp completes full processing including fact extraction."""
+    # Submit async retain with timestamp
+    response = await api_client.post(
+        f"/v1/default/banks/{test_bank_id}/memories",
+        json={
+            "items": [
+                {
+                    "content": "The quarterly meeting was held on January 30th 2026",
+                    "context": "meetings",
+                    "timestamp": "2026-01-30T11:45:00Z"
+                }
+            ],
+            "async": True
+        }
+    )
+    
+    assert response.status_code == 200, f"Expected 200, got {response.status_code}: {response.text}"
+    data = response.json()
+    assert data["success"] is True
+    assert data["async"] is True
+    operation_id = data["operation_id"]
+    
+    # Wait for async processing to complete (poll operation status)
+    max_wait_seconds = 30
+    poll_interval = 0.5
+    elapsed = 0
+    operation_completed = False
+    
+    while elapsed < max_wait_seconds:
+        response = await api_client.get(f"/v1/default/banks/{test_bank_id}/operations/{operation_id}")
+        if response.status_code == 200:
+            op_status = response.json()
+            if op_status.get("status") == "completed":
+                operation_completed = True
+                break
+            elif op_status.get("status") == "failed":
+                raise AssertionError(f"Operation failed: {op_status.get('error_message')}")
+        
+        await asyncio.sleep(poll_interval)
+        elapsed += poll_interval
+    
+    assert operation_completed, f"Async operation did not complete within {max_wait_seconds} seconds"
+    
+    # Verify memories were actually stored
+    response = await api_client.get(
+        f"/v1/default/banks/{test_bank_id}/memories/list",
+        params={"limit": 10}
+    )
+    assert response.status_code == 200
+    items = response.json()["items"]
+    assert len(items) > 0, "Should have stored memories after async processing"
@@ -156,7 +156,6 @@ class TestWorkerPoller:
            pool=pool,
            worker_id="test-worker-1",
            executor=mock_executor,
-            batch_size=10,
        )

        claimed = await poller.claim_batch()
@@ -177,8 +176,8 @@ class TestWorkerPoller:
            assert row["worker_id"] == "test-worker-1"

    @pytest.mark.asyncio
-    async def test_claim_batch_respects_batch_size(self, pool, clean_operations):
-        """Test that claim_batch respects the batch_size limit."""
+    async def test_claim_batch_respects_max_slots(self, pool, clean_operations):
+        """Test that claim_batch respects the max_slots limit."""
        from hindsight_api.worker import WorkerPoller

        # Create 10 pending tasks
@@ -196,12 +195,11 @@ class TestWorkerPoller:
                payload,
            )

-        # Claim with batch_size=3
        poller = WorkerPoller(
            pool=pool,
            worker_id="test-worker-1",
            executor=lambda x: None,
-            batch_size=3,
+            max_slots=3,  # Limit to 3 concurrent tasks
        )

        claimed = await poller.claim_batch()
@@ -238,11 +236,14 @@ class TestWorkerPoller:
            executor=mock_executor,
        )

-        # Execute the task
+        # Execute the task (fire-and-forget)
        task_dict = json.loads(payload)
        claimed_task = ClaimedTask(operation_id=str(op_id), task_dict=task_dict, schema=None)
        await poller.execute_task(claimed_task)

+        # Wait for background task to complete
+        completed = await poller.wait_for_active_tasks(timeout=5.0)
+        assert completed, "Task did not complete within timeout"
        assert len(executed) == 1

        # Verify task is marked as completed
@@ -283,11 +284,15 @@ class TestWorkerPoller:
            max_retries=3,
        )

-        # Execute (should fail and retry)
+        # Execute (should fail and retry) - fire-and-forget
        task_dict = json.loads(payload)
        claimed_task = ClaimedTask(operation_id=str(op_id), task_dict=task_dict, schema=None)
        await poller.execute_task(claimed_task)

+        # Wait for background task to complete
+        completed = await poller.wait_for_active_tasks(timeout=5.0)
+        assert completed, "Task did not complete within timeout"
+
        # Verify task is back to pending with incremented retry_count
        row = await pool.fetchrow(
            "SELECT status, retry_count, worker_id FROM async_operations WHERE operation_id = $1",
@@ -327,11 +332,15 @@ class TestWorkerPoller:
            max_retries=3,
        )

-        # Execute (should fail permanently)
+        # Execute (should fail permanently) - fire-and-forget
        task_dict = json.loads(payload)
        claimed_task = ClaimedTask(operation_id=str(op_id), task_dict=task_dict, schema=None)
        await poller.execute_task(claimed_task)

+        # Wait for background task to complete
+        completed = await poller.wait_for_active_tasks(timeout=5.0)
+        assert completed, "Task did not complete within timeout"
+
        # Verify task is marked as failed
        row = await pool.fetchrow(
            "SELECT status, error_message FROM async_operations WHERE operation_id = $1",
@@ -388,7 +397,6 @@ class TestWorkerPoller:
            pool=pool,
            worker_id="test-worker-1",
            executor=lambda x: None,
-            batch_size=10,
        )

        claimed = await poller.claim_batch()
@@ -440,7 +448,6 @@ class TestWorkerPoller:
            pool=pool,
            worker_id="test-worker-1",
            executor=lambda x: None,
-            batch_size=10,
        )

        claimed = await poller.claim_batch()
@@ -607,7 +614,6 @@ class TestConcurrentWorkers:
                pool=pool,
                worker_id=worker_id,
                executor=lambda x: None,
-                batch_size=5,  # Each worker tries to claim 5
            )
            claimed = await poller.claim_batch()
            workers_claimed[worker_id] = [task.operation_id for task in claimed]
@@ -680,7 +686,6 @@ class TestConcurrentWorkers:
            pool=pool,
            worker_id="new-worker",
            executor=lambda x: None,
-            batch_size=10,
        )

        claimed = await poller.claim_batch()
@@ -879,7 +884,6 @@ class TestDynamicTenantDiscovery:
            pool=pool,
            worker_id="test-worker-1",
            executor=lambda x: None,
-            batch_size=10,
            tenant_extension=mock_extension,
        )

@@ -946,7 +950,6 @@ class TestDynamicTenantDiscovery:
            pool=pool,
            worker_id="test-worker-1",
            executor=lambda x: None,
-            batch_size=10,
            tenant_extension=dynamic_extension,
        )

@@ -1008,7 +1011,6 @@ class TestDynamicTenantDiscovery:
            pool=pool,
            worker_id="test-worker-1",
            executor=lambda x: None,
-            batch_size=10,
        )

        claimed = await poller.claim_batch()
@@ -1017,3 +1019,198 @@ class TestDynamicTenantDiscovery:
        # All tasks should have schema=None (public)
        for task in claimed:
            assert task.schema is None
+
+
+async def test_worker_fire_and_forget_nonblocking(pool, clean_operations):
+    """
+    Test that worker continues polling while tasks run (fire-and-forget pattern).
+
+    This test verifies the FIX: With the old blocking behavior, the worker would
+    wait for all tasks in a batch to complete before claiming more. This test
+    would FAIL with the old code because tasks 3-4 wouldn't be claimed until
+    tasks 1-2 complete. With fire-and-forget, tasks 3-4 are claimed immediately.
+    """
+    from hindsight_api.worker.poller import WorkerPoller
+
+    task_started = {}  # operation_id -> Event (set when task starts)
+    task_canfinish = {}  # operation_id -> Event (wait before finishing)
+
+    async def blocking_executor(task_dict: dict):
+        op_id = task_dict["operation_id"]
+        # Signal that this task has started
+        started = asyncio.Event()
+        task_started[op_id] = started
+        started.set()
+
+        # Block until we're told to finish
+        finish = asyncio.Event()
+        task_canfinish[op_id] = finish
+        await finish.wait()
+
+    poller = WorkerPoller(
+        pool=pool,
+        worker_id="test-worker",
+        executor=blocking_executor,
+        poll_interval_ms=50,  # Fast polling
+        max_slots=10,
+        consolidation_max_slots=2,
+    )
+
+    bank_id = f"test-worker-{uuid.uuid4().hex[:8]}"
+
+    # Submit initial 2 tasks
+    task_ids = []
+    for i in range(2):
+        op_id = uuid.uuid4()
+        task_ids.append(str(op_id))
+        payload = json.dumps({"type": "test", "operation_type": "retain", "operation_id": str(op_id), "bank_id": bank_id})
+        await pool.execute(
+            """
+            INSERT INTO async_operations (operation_id, bank_id, operation_type, status, task_payload)
+            VALUES ($1, $2, 'retain', 'pending', $3::jsonb)
+            """,
+            op_id,
+            bank_id,
+            payload,
+        )
+
+    poll_task = asyncio.create_task(poller.run())
+
+    try:
+        # Wait for first 2 tasks to start executing (but not finish)
+        for i in range(100):  # Try for up to 1 second
+            if len(task_started) >= 2:
+                break
+            await asyncio.sleep(0.01)
+        assert len(task_started) == 2, f"Expected 2 tasks started, got {len(task_started)}"
+
+        # Verify tasks are in_flight
+        async with poller._in_flight_lock:
+            assert poller._in_flight_count == 2
+
+        # NOW submit 2 more tasks WHILE the first 2 are still running
+        for i in range(2):
+            op_id = uuid.uuid4()
+            task_ids.append(str(op_id))
+            payload = json.dumps({"type": "test", "operation_type": "retain", "operation_id": str(op_id), "bank_id": bank_id})
+            await pool.execute(
+                """
+                INSERT INTO async_operations (operation_id, bank_id, operation_type, status, task_payload)
+                VALUES ($1, $2, 'retain', 'pending', $3::jsonb)
+                """,
+                op_id,
+                bank_id,
+                payload,
+            )
+
+        # KEY ASSERTION: Worker should claim tasks 3-4 WITHOUT waiting for 1-2 to finish
+        # This would FAIL with the old blocking behavior
+        for i in range(100):  # Try for up to 1 second
+            if len(task_started) >= 4:
+                break
+            await asyncio.sleep(0.01)
+
+        assert len(task_started) == 4, (
+            f"Fire-and-forget FAILED: Expected 4 tasks started, got {len(task_started)}. "
+            "This means the worker blocked waiting for the first batch to complete."
+        )
+
+        # Verify all 4 tasks are in-flight
+        async with poller._in_flight_lock:
+            assert poller._in_flight_count == 4
+
+        # Clean up: allow all tasks to finish
+        for event in task_canfinish.values():
+            event.set()
+
+    finally:
+        # Ensure cleanup
+        for event in task_canfinish.values():
+            event.set()
+        await poller.shutdown_graceful(timeout=2.0)
+        try:
+            await asyncio.wait_for(poll_task, timeout=1.0)
+        except asyncio.CancelledError:
+            pass
+
+
+async def test_worker_slot_limits_enforced(pool, clean_operations):
+    """Test that worker respects max_slots and won't exceed the limit."""
+    from hindsight_api.worker.poller import WorkerPoller
+
+    tasks_started = set()
+    task_events = {}
+
+    async def controlled_executor(task_dict: dict):
+        op_id = task_dict["operation_id"]
+        tasks_started.add(op_id)
+        event = asyncio.Event()
+        task_events[op_id] = event
+        await event.wait()
+
+    poller = WorkerPoller(
+        pool=pool,
+        worker_id="test-worker",
+        executor=controlled_executor,
+        poll_interval_ms=50,
+        max_slots=3,  # Only allow 3 concurrent tasks
+        consolidation_max_slots=1,
+    )
+
+    # Submit 10 tasks
+    bank_id = f"test-worker-{uuid.uuid4().hex[:8]}"
+    for i in range(10):
+        op_id = uuid.uuid4()
+        payload = json.dumps({"type": "test", "operation_type": "retain", "operation_id": str(op_id), "bank_id": bank_id})
+        await pool.execute(
+            """
+            INSERT INTO async_operations (operation_id, bank_id, operation_type, status, task_payload)
+            VALUES ($1, $2, 'retain', 'pending', $3::jsonb)
+            """,
+            op_id,
+            bank_id,
+            payload,
+        )
+
+    poll_task = asyncio.create_task(poller.run())
+
+    try:
+        # Wait for slots to fill
+        for i in range(100):
+            if len(tasks_started) >= 3:
+                break
+            await asyncio.sleep(0.01)
+
+        # Should have claimed exactly 3 tasks (slot limit)
+        assert len(tasks_started) == 3
+
+        # Wait to ensure no additional tasks are claimed
+        for i in range(30):
+            await asyncio.sleep(0.01)
+        assert len(tasks_started) == 3, "Worker exceeded slot limit!"
+
+        # Release tasks one by one and verify remaining are claimed
+        completed = 0
+        while completed < 10 and len(tasks_started) < 10:
+            # Release the next batch
+            events_to_release = list(task_events.values())[completed:completed+3]
+            for event in events_to_release:
+                event.set()
+            completed += len(events_to_release)
+
+            # Wait for new tasks to be claimed
+            for i in range(100):
+                if len(tasks_started) >= min(completed + 3, 10):
+                    break
+                await asyncio.sleep(0.01)
+
+        assert len(tasks_started) == 10
+
+    finally:
+        for event in task_events.values():
+            event.set()
+        await poller.shutdown_graceful(timeout=2.0)
+        try:
+            await asyncio.wait_for(poll_task, timeout=1.0)
+        except asyncio.CancelledError:
+            pass
@@ -1,6 +1,6 @@
 [package]
 name = "hindsight-cli"
-version = "0.4.2"
+version = "0.4.5"
 edition = "2021"
 authors = ["Hindsight Team"]
 description = "A beautiful CLI for Hindsight - semantic memory system"
@@ -7,14 +7,14 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
 """  # noqa: E501


-__version__ = "0.0.7"
+__version__ = "0.4.5"

 # import apis into sdk package
 from hindsight_client_api.api.banks_api import BanksApi
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -489,7 +489,7 @@ class Configuration:
        return "Python SDK Debug Report:\n"\
               "OS: {env}\n"\
               "Python Version: {pyversion}\n"\
-               "Version of the API: 0.4.2\n"\
+               "Version of the API: 0.4.3\n"\
               "SDK Package Version: 0.0.7".\
               format(env=sys.platform, pyversion=sys.version)

@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -6,7 +6,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
@@ -5,7 +5,7 @@

    HTTP API for Hindsight

-    The version of the OpenAPI document: 0.4.2
+    The version of the OpenAPI document: 0.4.3
    Generated by OpenAPI Generator (https://openapi-generator.tech)

    Do not edit the class manually.
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Nicolò Boschi	823ab4b44e	doc: show embed page	2026-01-30 15:53:17 +01:00
Nicolò Boschi	fe88bdf704	Release v0.4.5 - Update version to 0.4.5 in all components - Python packages: hindsight-api, hindsight-dev, hindsight-all, hindsight-litellm, hindsight-embed - Python client: hindsight-clients/python - TypeScript client: hindsight-clients/typescript - Rust CLI: hindsight-cli - Control Plane: hindsight-control-plane - OpenClaw integration: hindsight-integrations/openclaw - Helm chart - Sync documentation to version-0.4	2026-01-30 14:55:50 +01:00
Nicolò Boschi	cbb8fc6723	fix: retain async with timestamp might fails (#253 )	2026-01-30 14:54:32 +01:00
Nicolò Boschi	c33b9b8bb2	Release v0.4.4 - Update version to 0.4.4 in all components - Python packages: hindsight-api, hindsight-dev, hindsight-all, hindsight-litellm, hindsight-embed - Python client: hindsight-clients/python - TypeScript client: hindsight-clients/typescript - Rust CLI: hindsight-cli - Control Plane: hindsight-control-plane - OpenClaw integration: hindsight-integrations/openclaw - Helm chart - Sync documentation to version-0.4	2026-01-30 13:05:07 +01:00
Nicolò Boschi	b364bc3402	fix: rename openclawd to openclaw (#252 ) * fix: rename openclawd to openclaw * fix: rename openclawd to openclaw * Revise OpenClaw documentation and remove dev section Updated the description of local memory for OpenClaw agents and removed the development section along with requirements and links.	2026-01-30 13:04:42 +01:00
Nicolò Boschi	35f0984b72	fix: retain async fails if timestamp is set (#251 ) * fix: retain async fails if timestamp is set * fix: rename openclawd to openclaw	2026-01-30 13:04:33 +01:00
Nicolò Boschi	5dc45194c9	sync docs	2026-01-30 11:47:38 +01:00
Nicolò Boschi	ff47814422	docs: improve openclawd integration docs - align with blog narrative - Fix XML tag: <hindsight-context> → <hindsight_memories> - Remove embedPort config option (not implemented in code) - Add default bankMission text to config docs - Add 'Why Auto-Recall?' section explaining conceptual advantage over tools - Add JSON format example showing metadata structure - Add 'Local-First Design' section emphasizing privacy/cost/ownership benefits - Update intro to highlight local-first and zero-cost aspects These changes better align the docs with the blog post's narrative about why auto-recall is better than tool-based memory and why local-first matters.	2026-01-30 11:47:03 +01:00
Nicolò Boschi	1ba70f81c8	sync docs to 0.4	2026-01-30 11:39:40 +01:00
Nicolò Boschi	fe15b5ec87	doc: openclawd	2026-01-30 11:28:29 +01:00
Nicolò Boschi	10e21f7302	changelog	2026-01-30 11:12:12 +01:00
Nicolò Boschi	7d3ac5ddb9	Release v0.4.3 - Update version to 0.4.3 in all components - Python packages: hindsight-api, hindsight-dev, hindsight-all, hindsight-litellm, hindsight-embed - Python client: hindsight-clients/python - TypeScript client: hindsight-clients/typescript - Rust CLI: hindsight-cli - Control Plane: hindsight-control-plane - OpenClawd integration: hindsight-integrations/openclawd - Helm chart - Sync documentation to version-0.4	2026-01-30 11:09:43 +01:00
Nicolò Boschi	f4f86e3842	fix: deadlock in worker polling (#250 ) * fix: deadlock in worker polling * fix: deadlock in worker polling * fixes	2026-01-30 11:09:29 +01:00
Nicolò Boschi	728ce13cea	fix: rename moltbot to openclawd (#246 ) * fix: rename moltbot to openclawd * fix * fix * fix: use single shared pg0 database for all banks + add default mission This commit fixes a critical database isolation issue and adds the default mission feature for the openclawd plugin. ## Changes: hindsight-embed: - Fixed daemon_client.py to use single shared database: pg0://hindsight-embed - Previously, each bank_id would create a separate pg0 instance (wrong!) - Now all banks share the same database with isolation via bank_id parameter - Updated README to clarify database architecture openclawd plugin (v0.0.5): - Added default bank mission describing OpenClawd's multi-channel assistant role - Added setBankMission() method to client - Integrated mission setting during plugin initialization - Added bankMission to plugin config schema with sensible default - Updated docs to explain shared database architecture ## Why this matters: Bank isolation should happen WITHIN the database (via separate tables/schemas), not via separate database instances. Using HINDSIGHT_EMBED_BANK_ID to create separate pg0 databases was architecturally wrong and caused confusion. * ci: rename moltbot to openclawd in workflows and release script - Updated build-moltbot-integration → build-openclawd-integration in test.yml - Updated release-moltbot-integration → release-openclawd-integration in release.yml - Updated all working directories from moltbot to openclawd - Updated artifact names from moltbot-integration to openclawd-integration - Added openclawd package.json to release.sh version bump script	2026-01-30 10:31:29 +01:00
Anton EvseevandClaude Opus 4.5	ecc590cb79	fix(docker): add retry logic for ML model downloads (#248 ) - Add 3 retries with exponential backoff (10s -> 20s -> 40s) - Set HF_HUB_DOWNLOAD_TIMEOUT=600 for longer timeout - Fixes transient network failures during HuggingFace downloads - Applied to both api-only and standalone stages Co-authored-by: Claude Opus 4.5 <[email protected]>	2026-01-30 09:42:16 +01:00
Nicolò Boschi	381c96c093	fix: improve doc on vertexai and mcp (#247 ) * fix: improve doc on vertexai and mcp * fix	2026-01-30 09:41:48 +01:00