ByteChef Launches Managed RAG Knowledge Base with Vector Search and OCR Support

ByteChef has introduced a built-in Knowledge Base feature that lets teams store and semantically search documents using retrieval-augmented generation (RAG). Users can upload files in formats including PDF, Markdown, Word, JSON, and plain text, with scanned documents handled via an OCR pipeline. The platform automatically parses, chunks, embeds, and indexes content into a pgvector-backed PostgreSQL database, with configurable chunk sizes and overlap settings per knowledge base. Retrieval is accessible through AI Agent tools, workflow actions, and a built-in search interface for testing queries before deployment. ByteChef manages the vector schema and indexing automatically, though users remain responsible for scheduling re-ingestion and manually handling upstream document deletions that the system cannot detect on its own.