What is RAG?

RAG stands for Retrieval‑Augmented Generation, a hybrid approach that combines a retrieval system with a generative language model. In a typical RAG pipeline, a user query first triggers a retrieval component—often a vector search over a large document corpus—that fetches the most relevant passages. These retrieved snippets are then fed as context into a generative model (e.g., GPT‑3 or BART), which produces a final answer conditioned on both the query and the retrieved content. This architecture improves factual accuracy and reduces hallucination because the model can reference real documents. RAG is widely used in question‑answering systems, knowledge‑base chatbots, and any application where up‑to‑date or domain‑specific information is critical. By decoupling retrieval from generation, RAG also allows for efficient scaling: the retrieval index can be updated independently of the generative model, enabling rapid incorporation of new data.