Ragas is an open-source toolkit to evaluate and monitor Retrieval-Augmented Generation applications. It offers automatic metrics to assess context relevance, precision/recall, and answer quality, can synthetically generate diverse evaluation datasets, and supports online monitoring to ensure production quality. It integrates with LangChain, LangSmith, and LlamaIndex, and is available via pip.