langbot-app/LangBot

Production-grade platform for building agentic IM bots - 生产级多平台智能机器人开发平台/ Agent、知识库编排、插件系统 / Bots for Discord / Slack / LINE / Telegram / WeChat(企业微信, 企微智能机器人, 公众号) / 飞书 / 钉钉 / QQ / Matrix e.g. Integrated with ChatGPT(GPT), DeepSeek, Dify, n8n, Langflow, Coze, Claude, Gemini, GLM, Ollama, SiliconFlow, Moonshot, openclaw / hermes agent, deerflow

LangBot – Open‑source platform for AI‑powered instant‑messaging bots

What it is – LangBot is a production‑grade, open‑source framework that lets you connect large language models (OpenAI, Anthropic, DeepSeek, local Ollama, etc.) to a wide range of instant‑messaging services (Discord, Slack, Telegram, WeChat, LINE, QQ, WeCom, Lark, DingTalk, KOOK, Matrix, …). It provides a web dashboard, a plugin ecosystem, and built‑in RAG / tool‑calling so you can build, debug, monitor, and ship multi‑turn conversational agents without writing platform‑specific glue code.

Core capabilities

Capability	What you get
Universal IM support	One codebase talks to >10 chat platforms; the dashboard lets you configure each bot’s credentials and routing.
Agent‑friendly architecture	Built‑in Model Context Protocol (MCP) server (`/mcp`) mirrors the HTTP API, enabling other AI agents to create/modify pipelines, plugins, and models programmatically.
RAG & tool calling	Integrated knowledge‑base connectors (Dify, Coze, n8n, Langflow, Deerflow, Weknora) and multi‑modal streaming output.
Production features	Access control, rate limiting, sensitive‑word filtering, real‑time monitoring (message volume, model calls, success rate), and robust exception handling.
Plugin ecosystem	Hundreds of community plugins, event‑driven extensions, and a marketplace (LangBot Space) that can be queried via MCP.
Web management panel	Browser UI for creating bots, wiring pipelines, managing keys, and viewing metrics – no YAML editing required.
Multi‑pipeline	Separate pipelines per bot or scenario, each with its own monitoring and error handling.

Supported LLM providers (as of the README)

OpenAI, Anthropic, DeepSeek, Google Gemini, xAI, Moonshot, Zhipu AI
Local models via Ollama or LM Studio
LLM‑ops platforms: Dify, SiliconFlow, Aliyun Bailian, Volc Engine, ModelScope, GiteeAI, 302.AI, Qiniu, etc.
GPU‑cloud platforms: CompShare, PPIO, ShengSuanYun

Quick start options

LangBot Cloud (hosted SaaS) – Zero‑setup, ready‑to‑use at https://space.langbot.app/cloud.

One‑line local launch (requires the uv package manager):

uvx langbot   # starts the server on http://localhost:5300

Docker Compose – Clone the repo, cd docker, then docker compose --profile all up -d.
One‑click deployments – Buttons for Zeabur and Railway; also docs for plain Docker, manual install, BTPanel, or Kubernetes.

A public demo is available at https://demo.langbot.dev/ (demo credentials provided).

Typical use cases

Scenario	How LangBot helps
Customer support	Deploy an AI agent on Slack/Discord/Telegram that answers queries using your internal knowledge base via RAG.
Internal tooling	Hook n8n or Dify workflows to WeCom/DingTalk for automated approvals, data look‑ups, or report generation.
Community moderation	AI‑driven content filtering and interactive assistance in QQ, Discord, or other group chats.
Cross‑platform presence	Write the bot once; manage all channels from a single dashboard, reducing duplication and operational overhead.

Development & extensibility

Skills directory (skills/) – Central place for plugin code, deployment scripts, and end‑to‑end tests; agents can point here to understand how to build or extend the system.
MCP server – Exposes the same API used by the web UI over a protocol designed for AI agents, enabling programmatic bot creation and pipeline manipulation.
AGENTS.md / CLAUDE.md – Documentation aimed at LLM agents describing architecture conventions and the rule that API changes must keep MCP and skills in sync.
llms.txt – Machine‑readable project metadata published on the website for LLM consumption.

Community & resources

Discord – active community for support and plugin sharing.
Docs – feature list, deployment guides, and API reference at https://langbot.app/docs.
Plugin marketplace – searchable via the LangBot Space web UI or MCP.
Roadmap – public at https://langbot.featurebase.app/roadmap.

TL;DR

LangBot is a ready‑to‑run, open‑source framework that turns any LLM into a multi‑platform instant‑messaging bot with a full‑featured web UI, monitoring, and an extensible plugin system. It’s geared toward developers and enterprises that need a single, maintainable codebase for AI agents across dozens of chat services.