WeChat Bot

WeChat Bot is a self-hosted AI agent for people who want automated replies across messaging apps and tools to examine their WeChat conversations. It builds on Wechaty and can use local models through Ollama or external AI services.

• Messaging replies: Connect WeChat, Telegram, WhatsApp or Lark to an AI backend, or use Pi as the agent behind replies.
• Reply controls: Restrict WeChat replies to approved contacts and groups, with mentions required in group chats; Telegram and Lark also support allowlists and reply triggers.
• Local WeChat data: Read and search cached chats, contacts, group members, favorites and Moments through OpenCLI wx-cli without an LLM.
• Conversation analysis: Get local statistics for a group or individual contact, or send recent message samples to a selected model for deeper analysis.

The MIT-licensed project is open source and runs on your own machine or server with Node.js, with Docker deployment also supported. Backends include Ollama, ChatGPT, Claude, DeepSeek, Kimi and self-hosted Dify. Ollama can keep model processing local; cloud backends send reply content or analysis samples to the selected provider and require provider credentials. Local statistics don't call an AI service.

Live messaging needs an internet connection and platform authentication. WhatsApp uses Meta's Cloud API and requires a publicly reachable HTTPS endpoint. WeChat's default Web protocol can trigger account warnings or bans, and its automatic reply path handles text messages only.