Retrieval Integrations

Vector stores for semantic search and RAG. Configure pgvector (built-in), Pinecone, Weaviate, or Qdrant. Query knowledge bases for chunks or full answers with citations.

Vector backends

pgvector — Default, built into Postgres. No extra setup.

Pinecone — Cloud. Create index, add API key to config.

Weaviate — Local (Docker) or cloud.

Qdrant — Local (Docker) or cloud.

Create vector indexes in Dashboard → Knowledge → Vector indexes. Assign to a knowledge base.

RAG API

POST /retrieval/query — Semantic search. Returns chunks only.

Body: knowledge_base_id, query, top_k

POST /rag/run — Full RAG. Retrieve + generate answer with citations.

Body: knowledge_base_id, provider_model_id (chat model), question

Try RAG

Dashboard → Knowledge — Add a knowledge base, add sources, upload documents, run ingestion. Then use the Try RAG modal to ask questions and see answers with citations.