RAG (Retrieval-Augmented Generation)

A technique where an AI model retrieves relevant information from a knowledge source before generating an answer.

RAG combines a retrieval step (searching a document store, usually via a vector database, for passages relevant to a question) with a generation step (an LLM writing an answer grounded in those retrieved passages). This lets an AI system answer questions about specific, current, or private information it wasn't trained on, and cite its sources — rather than relying only on what the model memorized during training.