Glossary
What is RAG?
Giving the AI your own documents to pull answers from, so it responds from real sources instead of memory alone.
RAG — retrieval-augmented generation — is the trick that lets an AI answer from your documents instead of only its training. Before answering, the system searches your files for the relevant passages and pastes them into the prompt, so the model responds from real, quotable sources rather than memory.
The model reads your sources at answer time — no retraining.
RAG is how most "chat with your data" tools work, and it's the main cure for hallucination on private or fresh information. It has limits too: an answer is only as good as what search finds, so how you chunk and index your documents matters.
Related terms
Appears in these lessons
See it live
Build a "chat with your docs" in miniature
See retrieval feed the answer, step by step — free in the app.