LangChain-Chatchat is an open-source RAG and agent application designed for local and offline deployment, with particular attention to Chinese-language scenarios and open-source models. Originally known as Langchain-ChatGLM, it provides a complete knowledge-base question-answering workflow built around local model serving and LangChain-style application components.

Its RAG pipeline loads documents, extracts and splits text, creates embeddings, retrieves relevant passages, and supplies them as context to an LLM. The project supports mainstream open models, embedding models, rerankers, vector databases, and deployment frameworks such as Xinference, Ollama, LocalAI, and FastChat. Newer versions also add stronger agent capabilities, hybrid retrieval approaches, multimodal interactions, and a multi-session web interface.

LangChain-Chatchat is aimed at developers and organizations that want private knowledge assistants without depending entirely on hosted APIs. It can operate with fully open-source local components while still supporting OpenAI-compatible services when desired, making it useful for internal document Q&A, private RAG systems, experimental agents, and deployments where data residency or offline operation is important.