RAGAS

RAGAS is an open-source evaluation framework for retrieval-augmented generation systems. It decomposes RAG quality into four primary metrics: faithfulness (does the answer stick to the retrieved context), answer relevance (does the answer address the question), context precision (how useful were the retrieved documents), and context recall (were all relevant documents retrieved). Faithfulness and answer relevance are scored using LLM-as-judge internally, while context precision and recall compare retrieval results against reference annotations. RAGAS is the standard starting point for teams that want quantitative signals on RAG pipelines rather than eyeballing outputs, and it plugs into most popular RAG stacks.