langbot

LangBot provides a self-hostable platform for connecting large language models and agent workflows to instant-messaging services. It supports multi-turn conversations, streaming responses, multimodal interaction, tool calling, and built-in RAG/knowledge-base capabilities. Bots can be deployed across Discord, Telegram, Slack, LINE, QQ, WeChat, WeCom, Lark, DingTalk, KOOK, Satori, email, and Matrix, with Matrix bridging to additional chat networks. The platform integrates with model and agent services including OpenAI, Anthropic, DeepSeek, Google Gemini, xAI, Moonshot, Zhipu AI, Ollama, LM Studio, Dify, MCP, SiliconFlow, and several gateways or GPU platforms. It also connects with automation and orchestration tools such as n8n, Langflow, Coze, and Deerflow. LangBot includes a browser-based administration panel for configuration, bot management, pipelines, models, API/MCP access, and operational monitoring. Production-oriented features listed in the repository include access control, rate limiting, sensitive-word filtering, exception handling, session and message monitoring, and multi-pipeline support. An event-driven plugin system and plugin marketplace allow extending bot behavior. The built-in MCP server exposes management capabilities over HTTP for coding agents, while the repository includes skills and agent guidance for plugin development, testing, deployment, and core development. Deployment options include LangBot Cloud, a one-line `uvx langbot` launch, Docker Compose, Docker, manual installation, Kubernetes, Zeabur, Railway, and one-click panel deployment. The project is written for Python 3.10 through 3.13.