All ClawAll Claw
Back to Catalog
WeClaw logo
tools

WeClaw

Open-source WeChat AI Agent Bridge (ClawBot) that connects WeChat to any AI coding agents like Claude, Codex, Gemini, Cursor, Kimi, and OpenClaw for seamless chatting, commands, and proactive messaging.

Overview

WeClaw is an open-source WeChat AI Agent Bridge (also known as ClawBot) that lets you interact with powerful AI coding agents directly inside WeChat. Built in Go, it automatically detects installed agents, routes messages via chat commands or aliases, and supports multiple communication modes for fast, persistent sessions.

It enables chatting with agents like Claude, Codex, Gemini, Kimi, Cursor, OpenCode, and OpenClaw without leaving WeChat, plus proactive sending and background operation.

Features

  • Multi-Agent Support: Seamlessly connects to Claude, Codex, Gemini, Kimi, Cursor, OpenCode, OpenClaw, and more.
  • Flexible Agent Modes: ACP (JSON-RPC over stdio) — fastest, persistent sessions; CLI — spawns processes per message with session support; HTTP — OpenAI-compatible chat completions API.
  • Rich Chat Commands: hello to default agent; /codex <prompt>, /cc <prompt> (aliases like /claude, /cx, /gm, /oc for OpenClaw); /status, agent switching, and more.
  • Proactive Messaging: Send messages via CLI (weclaw send) or HTTP API (POST /api/send).
  • Background & Service Mode: Run as daemon (start/stop/status), with launchd (macOS) and systemd (Linux) support.
  • Auto-Detection & Login: Scans QR code on first run for WeChat login; auto-detects local agents.
  • Docker & CLI Support: Easy containerized deployment and one-line install.
  • Config Persistence: Stores settings in ~/.weclaw/config.json with environment variable overrides.

Use Cases

  • Chat with your favorite AI coding assistant (e.g., Claude or Cursor) directly in WeChat on mobile or desktop.
  • Build WeChat-based AI workflows, automation, or personal assistants.
  • Integrate multiple agents and switch instantly via simple commands.
  • Run headless/background for always-on agent access with proactive notifications.
  • Develop or test AI agents using WeChat as the front-end interface.

Installation & Getting Started

One-line install:

curl -sSL https://raw.githubusercontent.com/fastclaw-ai/weclaw/main/install.sh | sh

Start the bridge:

weclaw start

Scan the QR code with WeChat on first run. Agents are auto-detected.

Alternative methods: go install github.com/fastclaw-ai/weclaw@latest or Docker.

Full setup, config examples, service installation, and commands are in the GitHub README.

Technical Details

  • Language: Go (primary), with shell scripts for install/services.
  • License: MIT.
  • Repository: https://github.com/fastclaw-ai/weclaw (released March 2026, active development).
  • Stars: ~175 (as of March 2026).

Related Projects

Note: There are separate projects using similar names (e.g., commercial WeClawBot desktop AI app, or tomeai/weclaw MCP Gateway). This entry focuses on the open-source fastclaw-ai/weclaw bridge.

Tags

wechatai-agentclawbotopenclawclaudegeminigolangclibridgeautomationweixin