Langchain-Chatchat

Why we included this project

Langchain-Chatchat gives you a self-hosted question-answering system over your own documents, with the full RAG pipeline handled for you: file loading, text splitting, embedding, vector search, and prompt assembly all run against a local model like ChatGLM, Qwen, or Llama, so nothing leaves your infrastructure. It also supports agent-style tool calling, conversations with search engines and databases, and a Streamlit web UI for managing multiple sessions. The project is a good fit for teams that need Chinese-language support and want to switch between model backends such as Ollama, Xinference, FastChat, or LocalAI without reworking the application. Treat it as a ready-made starting point for a private document assistant rather than a library to embed inside another product.