LangBot is an open-source, production-grade platform for building AI-powered instant messaging bots, aimed at developers and teams who need to connect large language models to chat platforms such as Discord, Telegram, Slack, LINE, and WeChat.

What it is

LangBot is a Python project, licensed under Apache-2.0, that links Large Language Models to instant messaging platforms so that teams can build agents which converse, execute tasks, and connect to existing workflows. It sits in the AI and machine learning tooling ecosystem, filed under AI Development Platforms, and it ships as a single codebase covering Discord, Telegram, Slack, LINE, QQ, WeChat, WeCom, Lark, DingTalk, KOOK, and Matrix. The project carries 17,880 stars and 1,598 forks, with documentation translated into nine languages, a plugin market, and a roadmap published alongside the source.

The concrete problem it solves is per-platform bot plumbing and hand-edited configuration. Rather than writing a separate bot for each messenger and maintaining its configuration in YAML files, LangBot provides one runtime with a browser-based management panel, so bots are configured, managed, and monitored without editing YAML. It also supplies the production concerns that a naive bot script omits: access control, rate limiting, sensitive word filtering, comprehensive monitoring, and exception handling, all applied across the pipelines that connect models to chat surfaces.

Key capabilities

• AI conversations and agents with multi-turn dialogues, tool calling, multi-modal support, and streaming output.
• Built-in RAG knowledge base with deep integration to Dify, Coze, n8n, Langflow, Deerflow, and Weknora.
• Universal IM platform support from one codebase: Discord, Telegram, Slack, LINE, QQ, WeChat, WeCom, Lark, DingTalk, and KOOK.
• Production features including access control, rate limiting, sensitive word filtering, comprehensive monitoring, and exception handling.
• Plugin ecosystem with hundreds of plugins, event-driven architecture, component extensions, and MCP protocol support.
• Web management panel for configuring, managing, and monitoring bots, with no YAML editing required.
• Multi-pipeline architecture that runs different bots for different scenarios under the same deployment.
• Model support spanning ChatGPT (GPT), DeepSeek, Claude, Gemini, GLM, Ollama, SiliconFlow, Moonshot, openclaw, hermes agent, and deerflow.

Who uses it and how

• Teams shipping one bot to many messengers use the single codebase to reach Discord, Telegram, Slack, LINE, and WeChat at once.
• Enterprises rely on the built-in access control, rate limiting, and sensitive word filtering to expose agents to real user populations.
• Knowledge-driven deployments use the built-in RAG knowledge base or attach an external Dify, Coze, or Langflow agent as the answer engine.
• Automation teams route conversations through n8n so that chat messages trigger wider workflows.
• Operators running several distinct bots for separate scenarios use the multi-pipeline architecture instead of separate installations.

Getting started

The quickest launch is the one-line command uvx langbot, which requires uv and serves the panel at http://localhost:5300. Alternatives include cloning the repository and running docker compose --profile all up -d from LangBot/docker, the hosted LangBot Cloud for zero deployment, one-click templates on Zeabur and Railway, and documented Docker, manual, BTPanel, and Kubernetes paths.

How it compares

No comparable open-source tools are named in the registry entry, so LangBot stands alone in this registry. It also does not list paid products that it replaces, so no licence, hosting, or cost contrast can be drawn against commercial alternatives here.

When to use it — and when not to

Choose LangBot when the goal is one maintained runtime across many messaging platforms with a panel instead of YAML files. A self-hoster must still operate the deployment itself, whether through uv, Docker Compose, manual installation, or Kubernetes, and must configure each platform integration and model provider individually. Teams that only need a single messenger bot with no production controls will find the platform larger than the task, and the repository currently carries 119 open issues, so expect active churn rather than a frozen surface.

uvx langbot

git clone https://github.com/langbot-app/LangBot
cd LangBot/docker
docker compose --profile all up -d