What is this?
An open-source, production-grade platform for building AI-powered instant messaging bots: it connects LLMs to chat platforms so you can ship agents that converse, execute tasks, and plug into your workflows.
An open-source, production-grade platform for building AI-powered instant messaging bots: it connects LLMs to chat platforms so you can ship agents that converse, execute tasks, and plug into your workflows.Is it a hosted service I sign up for, or something I self-host?Both: there is a zero-deployment LangBot Cloud, but the main path is self-hosting — you deploy the platform and manage, configure, and monitor your bots from a browser-based web panel, no YAML editing required.
Both: there is a zero-deployment LangBot Cloud, but the main path is self-hosting — you deploy the platform and manage, configure, and monitor your bots from a browser-based web panel, no YAML editing required.How do I get a bot running?Try the public live demo first (sign in with [email protected] / langbot123456 — keep sensitive data out of it). To self-host: a one-line launch that requires uv, Docker Compose, or one-click cloud deploy; Docker, Manual, BTPanel, and Kubernetes are listed as further options.
Try the public live demo first (sign in with [email protected] / langbot123456 — keep sensitive data out of it). To self-host: a one-line launch that requires uv, Docker Compose, or one-click cloud deploy; Docker, Manual, BTPanel, and Kubernetes are listed as further options.Am I locked into one AI provider?No — OpenAI, Anthropic, DeepSeek, Gemini, xAI, Moonshot, and Zhipu AI are built in, local models run through Ollama and LM Studio, and there are gateways like SiliconFlow and 302.AI.
No — OpenAI, Anthropic, DeepSeek, Gemini, xAI, Moonshot, and Zhipu AI are built in, local models run through Ollama and LM Studio, and there are gateways like SiliconFlow and 302.AI.Is it really ready for production, or more of a toy?Production is the design goal: access control, rate limiting, sensitive word filtering, comprehensive monitoring, and exception handling are built in, and the project is described as trusted by enterprises.
Production is the design goal: access control, rate limiting, sensitive word filtering, comprehensive monitoring, and exception handling are built in, and the project is described as trusted by enterprises.What do people actually build with it?The README calls out customer-support agents answering from a knowledge base on Slack/Discord/Telegram, internal n8n/Dify workflows wired into WeCom/DingTalk, AI moderation for QQ/Discord communities, and one bot present across all platforms.
The README calls out customer-support agents answering from a knowledge base on Slack/Discord/Telegram, internal n8n/Dify workflows wired into WeCom/DingTalk, AI moderation for QQ/Discord communities, and one bot present across all platforms.Which chat platforms can a bot live on?Discord, Telegram, Slack, LINE, QQ, WeChat, WeCom, Lark, DingTalk, KOOK, Matrix, Satori, and Email — and Matrix bridges to Signal, WhatsApp, Messenger, iMessage, Google Chat, IRC, XMPP, and more.
Discord, Telegram, Slack, LINE, QQ, WeChat, WeCom, Lark, DingTalk, KOOK, Matrix, Satori, and Email — and Matrix bridges to Signal, WhatsApp, Messenger, iMessage, Google Chat, IRC, XMPP, and more.Can I run several different bots for different scenarios?Yes — the multi-pipeline architecture lets different bots serve different scenarios, each with its own monitoring and exception handling.
Yes — the multi-pipeline architecture lets different bots serve different scenarios, each with its own monitoring and exception handling.Can I extend it with plugins?There is a plugin ecosystem with hundreds of plugins, an event-driven architecture, and component extensions, plus MCP protocol support. Plugins, MCP servers, and skills are distributed through the LangBot Space marketplace.
There is a plugin ecosystem with hundreds of plugins, an event-driven architecture, and component extensions, plus MCP protocol support. Plugins, MCP servers, and skills are distributed through the LangBot Space marketplace.Can coding agents like Claude Code operate LangBot for me?Yes, it is agent-friendly by design: a built-in MCP server at /mcp mirrors the HTTP API so an agent can manage bots, pipelines, plugins, and models with your API key, and the repo ships a skills/ directory plus AGENTS.md describing architecture and conventions.
Yes, it is agent-friendly by design: a built-in MCP server at /mcp mirrors the HTTP API so an agent can manage bots, pipelines, plugins, and models with your API key, and the repo ships a skills/ directory plus AGENTS.md describing architecture and conventions.How does the knowledge base (RAG) side work?RAG is built in, with deep integration to Dify, Coze, n8n, Langflow, Deerflow, and Weknora, so bots can answer questions from your own documents.
RAG is built in, with deep integration to Dify, Coze, n8n, Langflow, Deerflow, and Weknora, so bots can answer questions from your own documents.