Ask anything about your documents

Your queries are answered using RAG — retrieving relevant chunks from uploaded documents via pgvector, then generating answers with source citations.