RAG (Retrieval-Augmented Generation) is a technique where you fetch relevant documents and inject them into Claude's context before asking a question.

Why RAG?

• Gives Claude access to private, proprietary, or up-to-date knowledge
• Reduces hallucination by grounding answers in real documents
• Cheaper than fine-tuning

Basic Pipeline

1. Embed documents into a vector database
2. Retrieve relevant chunks based on user query
3. Inject chunks into Claude's context window
4. Claude answers using retrieved context

Reference:

RAG with Claude guide

https://docs.anthropic.com/en/docs/build-with-claude/retrieval-augmented-generation