Retrieval-Augmented Generation (RAG)

RAG combines a language model with document retrieval. Instead of relying only on what the model learned during training, a RAG workflow fetches relevant information from your documents and passes it to the model alongside the question.

Typical use cases include:

• Answering questions about internal documentation
• Building FAQ systems from product or support content
• Querying large document collections in natural language
• Grounding model responses in up-to-date or domain-specific data

How RAG works

A RAG pipeline has three steps: store documents as vectors, retrieve the most relevant passages for a question, and inject them into the prompt.

• What is RAG?: How retrieval-augmented generation works and when to use it.

Build a RAG pipeline

• Create a vector store: Split documents into chunks and store them as embeddings.
• Retrieve context: Find the most relevant passages for a user question.
• Inject context into a prompt: Add retrieved passages to the model prompt.

Tutorial

Follow a complete example that builds a product FAQ assistant using a RAG pipeline:

• Product FAQ assistant