Summary

This project provides a bridge to connect WeChat direct messages with any ACP-compatible AI agent, such as Claude, Codex, or Gemini. It targets users who wish to automate or integrate AI capabilities into their WeChat communications, offering a seamless way for AI agents to receive and respond to messages.

How It Works

The system logs into WeChat using the iLink bot API, continuously polls for incoming 1:1 messages, and forwards them to a specified ACP-compatible agent via standard input/output (stdio). Agent replies are then formatted and sent back to the WeChat user. Each WeChat user is managed with a dedicated ACP session and subprocess, ensuring message processing is handled serially per user.

Quick Start & Requirements

To start, run npx wechat-acp --agent <preset|command>. Key requirements include Node.js 20+, a WeChat environment supporting the iLink bot API, and an ACP-compatible AI agent accessible locally or via npx. The initial setup involves a terminal-based QR code login process to authenticate the iLink bot.

Highlighted Details

• Features terminal-rendered QR code login for WeChat.
• Manages individual ACP agent sessions and subprocesses for each WeChat user.
• Includes built-in agent presets like copilot, claude, gemini, qwen, codex, and opencode.
• Supports custom raw agent commands and configuration files.
• Offers a background daemon mode for continuous operation.
• Handles direct messages exclusively, ignoring group chats.
• Automatically approves permission requests originating from the AI agent.

Maintenance & Community

The provided README does not contain specific details regarding notable contributors, sponsorships, community channels (like Discord or Slack), or a public roadmap.

Licensing & Compatibility

The project is released under the MIT license, which is generally permissive for commercial use and integration into closed-source projects. However, users should be mindful of the terms of service for the underlying WeChat iLink bot API.

Limitations & Caveats

Currently, the bridge only supports direct messages, excluding group chats. Agent communication is strictly limited to subprocess stdio. Permission requests from agents are automatically approved, and some preset agents may necessitate separate authentication steps before they can function correctly.