Skip to main content

Vector search entries with full content and auto-embedding (SECONDARY)

POST 
/api/v1/vector/entries

Secondary vector search endpoint - Returns full entry objects with S3 content.

This is the COMPREHENSIVE search method that fetches complete entry content from S3. Use this when you need full documents, not just snippets.

Auto-Embedding Support:

  • Provide query (text) for automatic embedding generation
  • OR provide query_vector (pre-computed embeddings)
  • If both provided, query_vector takes precedence

Flow:

  1. If query provided: Generate embeddings using configured provider
  2. Query vector is sent to Qdrant
  3. Matching chunks are grouped by entry_id
  4. Entry records are fetched from PostgreSQL
  5. Full content is retrieved from S3 (parallel batch operation)
  6. Results combine entries with their matching chunks

Performance:

  • With query_vector: 200-500ms
  • With query (auto-embed): 700ms-2.5s (includes embedding generation)

Use cases:

  • Retrieving full documents
  • Document-level results
  • When you need complete content
  • Lower-throughput, higher-quality scenarios

Request​

Responses​

Successful Response