Home › Glossary › RAG (Retrieval-Augmented Generation)

Last updated: 2026-04-18

What is RAG (Retrieval-Augmented Generation)?

Technique where the agent fetches relevant documents from a vector store or search index before answering, then grounds its response in them. Reduces hallucination and lets agents work over large corpora the LLM never saw during training.

See also

• /openclaw/skills-guide/

← Back to the full AI agent glossary.