A cross-platform personal assistant agent, WeClaw, addresses the need for seamless task planning and execution by bridging mobile QQ and Windows PC capabilities. It targets users and developers seeking automated, natural language-driven workflows across devices, offering customizable integrations with platforms like WhatsApp, Discord, and Telegram.

How It Works

WeClaw utilizes a multi-agent architecture for parallel task decomposition and execution, intelligently routing tasks via ReAct or ReCAP strategies based on complexity. It incorporates advanced context engineering, including compression and filesystem support, to mitigate context overflow issues. The system ensures stable long-running pipelines through secure sandboxing and asynchronous execution, supported by a unified skills integration mechanism and robust task state management for multi-turn automation.

Quick Start & Requirements

The recommended startup method is via the Web Console: execute python entry_console.py or python -m angel_console, then access http://127.0.0.1:7788 in a web browser. Essential prerequisites include environment variables for LLM API keys (e.g., LLM_API_KEY is required) and optional search APIs. QQ integration necessitates obtaining BOTPY_APPID and BOTPY_SECRET from the Tencent QQ Open Platform (https://q.qq.com/#/). Credentials can be managed via a local local_secrets.yaml file.

Highlighted Details

• Agent Console: A graphical interface providing unified operations such as chat with SSE streaming and ReAct trace, voice input with local transcription, task search, multi-channel support (Web, CLI, QQ, Discord), scheduled tasks, skills management, model configuration, and billing/audit features.
• Integrated MCP Stack: Offers local and remote MCP client runtime, including discovery, configuration, and management functionalities.
• Cross-Platform Collaboration: Facilitates seamless interaction and task execution between mobile QQ and Windows PC environments.
• Multi-Agent System: Enables parallel decomposition and execution of complex tasks for enhanced efficiency and capability.

Maintenance & Community

The project was open-sourced on February 3, 2026. The provided README does not detail specific core contributors, sponsorships, or community channels such as Discord or Slack.

Licensing & Compatibility

WeClaw is released under the MIT license. This permissive license generally allows for commercial use and integration within closed-source applications without significant restrictions.

Limitations & Caveats

As a recently open-sourced project (February 2026), WeClaw may be subject to rapid development and potential API changes. Certain functionalities, particularly QQ integration, require external registration and configuration steps. Voice input relies on browser recording and local transcription, which may introduce platform-specific dependencies.