<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> WeClaw bridges WeChat with AI agents like Claude, Codex, and Gemini, enabling AI-powered communication within the popular messaging platform. It targets users seeking to automate or enhance WeChat interactions using AI, offering benefits such as intelligent message handling, proactive communication, and media integration.

How It Works

The project employs three primary agent interaction modes: ACP (fastest, long-running subprocess via JSON-RPC), CLI (spawns a new process per message), and HTTP (OpenAI-compatible API). WeClaw automatically detects and prioritizes ACP agents. It intelligently processes AI responses, converting markdown to plain text for WeChat, stripping code fences, and handling embedded image URLs by downloading and re-uploading them to WeChat's CDN.

Quick Start & Requirements

• Installation: A one-line curl command is provided for quick setup: curl -sSL https://raw.githubusercontent.com/fastclaw-ai/weclaw/main/install.sh | sh. Alternative methods include Go installation (go install github.com/fastclaw-ai/weclaw@latest) and Docker.
• First Run: Execute weclaw start. This prompts a QR code for WeChat login, auto-detects installed AI agents, and saves configuration to ~/.weclaw/config.json. Additional WeChat accounts can be added via weclaw login.
• Prerequisites: A functional WeChat client is necessary. For ACP and CLI agent modes, the respective AI agent binaries must be installed and accessible.

Highlighted Details

• Chat Commands: Supports intuitive chat commands for interacting with specific agents (e.g., /codex write a function), switching default agents (e.g., /claude), and checking status (/status).
• Media Handling: Capable of sending and receiving images, videos, and files. It automatically processes markdown image URLs by downloading and uploading them to WeChat's CDN.
• Proactive Messaging: Enables sending messages, images, and files to WeChat users programmatically via a CLI interface (weclaw send) or an HTTP API.

Maintenance & Community

The provided README does not detail specific contributors, sponsorships, or community channels like Discord or Slack.

Licensing & Compatibility

• License: MIT.
• Restrictions: The project is explicitly designated "For personal learning only, not for commercial use," which overrides standard MIT license permissions for commercial applications.

Limitations & Caveats

• Commercial Use: Commercial use is strictly prohibited by the project's stated purpose.
• Docker Agent Support: ACP and CLI agents require their binaries to be mounted into the Docker container or included in a custom image, whereas HTTP agents function out-of-the-box.
• Safety Flags: Certain CLI agents may require flags like --dangerously-skip-permissions to bypass interactive prompts, which disables safety checks and carries inherent risks.