Ragas

Ragas is an open-source framework, created by Shahul ES and Jithin James under the ExplodingGradients project, that provides automated metrics for evaluating retrieval-augmented generation (RAG) systems, including faithfulness, answer relevancy, context precision, and context recall. It also generates synthetic evaluation datasets and supports production monitoring, and is integrated by LangChain’s LangSmith and LlamaIndex.