If you're building document RAGRetrieval-augmented generation — fetching relevant documents first and pasting them into the model's context so it answers from your data instead of memory.Full definition → or enterprise search, this gives you layout-aware OCR — bounding boxes, block types, and per-block confidence scores you can use for chunkingSplitting documents into passages small enough to embed and retrieve individually — the step that quietly determines whether retrieval works at all.Full definition → and citation — across 170 languages, and it runs in a single self-hostable container so documents never leave your infrastructure.
Terms in this piece · Glossary
RAG — Retrieval-augmented generation — fetching relevant documents first and pasting them into the model's context so it answers from your data instead of memory.
chunking — Splitting documents into passages small enough to embed and retrieve individually — the step that quietly determines whether retrieval works at all.
