Delete Source
Knowledge & Memories
Delete Knowledge
Permanently delete one or more knowledge sources by ID.
POST
Delete Source
When to use it
- Cleanup – remove outdated documents
- Right-to-deletion – fulfill GDPR/privacy requests for specific items
- Reprocessing – delete then re-ingest after changing metadata schema or content
DELETE /memories/delete_memory. To remove an entire tenant, use DELETE /tenants/delete.
Endpoint
- Auth: Bearer token
- Idempotency: Re-sending a previously-deleted ID returns
deleted: falsefor that item, but the request as a whole still returns200. - Async: No
Why POST and not DELETE? This endpoint accepts a JSON body with an array of IDs. Some HTTP infrastructure strips bodies from
DELETE requests, so HydraDB uses POST to ensure the bulk payload is reliably transmitted.Example
Request parameters
Response
Behavior notes
Partial success is possible. A request can return
success: true while some individual items fail. Always inspect results[] to confirm each deletion outcome – don’t rely on the top-level success flag alone.Cascade scope. Deleting a knowledge source removes its chunks, embeddings, and any graph entities/relations exclusively associated with it. Entities shared with other sources are retained.
Related endpoints
- Browse before deleting: List data – confirm what you’re about to remove
- Delete a single memory: Delete memory
- Delete an entire tenant: Delete tenant
Errors
Common codes:400 INVALID_PARAMETERS, 404 TENANT_NOT_FOUND, 422 VALIDATION_ERROR. See Error Responses for the full list.Authorizations
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Body
application/json