Connects AI coding agents (Claude Code, Codex, OpenCode) to messaging platforms like Slack, WhatsApp, Discord, Telegram, and Feishu, enabling developers to manage and interact with AI coding tasks remotely from any device, including mobile. It simplifies multi-agent workflows by allowing users to switch between agents and autonomous levels simply by renaming channels, offering a flexible and accessible development environment.

How It Works

HandClaw acts as a bridge, connecting AI coding agents like Claude Code, Codex, and OpenCode to popular messaging platforms such as Slack, WhatsApp, Discord, Telegram, and Feishu. Its core design leverages channel management for agent control: renaming a channel dynamically switches the active AI agent and its autonomous level (e.g., l0-claude-repo1). This approach facilitates multi-round conversations and allows users to manage projects and interact with AI agents from any device, including mobile phones, without being tied to a desktop environment.

Quick Start & Requirements

Installation requires Node.js (version 22+) and pnpm. The primary installation command is npm install -g handclaw@latest. A crucial prerequisite is the separate installation and setup of AI coding agents (Claude Code, Codex, OpenCode), as HandClaw only provides the integration layer. The handclaw onboard --install-daemon command assists with initial setup. Official documentation links are available for getting started and agent configuration.

Highlighted Details

• Real-time notifications upon task completion.
• Persistent switching between 'plan' and 'build' modes via commands like !code switch plan.
• Ability to halt running code CLI processes immediately using the !stop command.
• Streamlined project management through channel renaming, enabling dynamic agent and autonomous level changes.
• Support for multiple AI agents and messaging platforms, enabling flexible remote development.

Maintenance & Community

HandClaw is a fork of the OpenClaw project. Community support is available via a Discord server.

Licensing & Compatibility

The project is released under the MIT license, which generally permits commercial use and integration with closed-source projects.

Limitations & Caveats

A significant caveat is the mandatory separate installation and configuration of AI coding agents (Claude Code, Codex, OpenCode). The project's live demo URL is currently a placeholder, and the specific implications of the different 'autonomous levels' (l0, l1, l2) may require further investigation.