← Back to Glossary
RAG (Retrieval-Augmented Generation)
Retrieval-Augmented Generation
Before answering, the system looks up relevant information from your own data — then hands that to the LLM so it answers using real facts instead of guessing from memory.
Why it exists
LLMs have knowledge cutoffs and hallucinate obscure facts. RAG grounds responses in private, up-to-date document stores.
How it works
- User query comes in. 2) Retriever searches vector/keyword store for top chunks. 3) Chunks injected into LLM prompt. 4) Generator writes answer.
Related terms
Where you'll use this
roadmap
project
interview