clawatar

What This Skill Does

Clawatar transforms your AI agent from a static text interface into a fully animated, 3D interactive character. By integrating a web-based VRM (Virtual Reality Model) viewer, this skill allows your agent to communicate using body language, facial expressions, and real-time lip-synced audio. It supports a massive catalog of 162 animations, allowing the agent to express emotions like happiness, sadness, or surprise, and perform complex actions such as waving, thinking, or dancing. The viewer operates via a WebSocket server, allowing your local OpenClaw agent to push visual commands instantly. It is highly customizable, supporting custom VRM model imports, background scene changes, and professional TTS integration via ElevenLabs.

Installation

To install Clawatar, ensure you have Node.js installed on your system. Run the following commands in your terminal:

1. Install via the OpenClaw CLI: clawhub install openclaw/skills/skills/dongping-chen/clawatar
2. Navigate to the directory: cd ~/.openclaw/workspace/clawatar
3. Install dependencies: npm install
4. Start the viewer: npm run start Once running, the avatar interface is accessible at http://localhost:3000. You can drag and drop your own .vrm file directly into the browser or configure a persistent model URL in the clawatar.config.json file.

Use Cases

• Virtual VTubing: Create a personality-driven stream where the AI acts as a host.
• Interactive Companions: Give your agent a physical presence for more engaging and empathetic human-computer interaction.
• Presentations: Use the agent to demonstrate concepts with gestures and stage movements.
• Education: Create a 3D tutor that reacts to user progress with nods, head shakes, and celebratory dances.

Example Prompts

1. "OpenClaw, launch my Clawatar and wave to me. I want to test the lip-syncing for my presentation script."
2. "I'm feeling happy today, can you make the avatar look joyful and perform the 'Ymca Dance' while we talk?"
3. "Switch the avatar to the Sakura Garden background and perform a thinking animation while you process my request."

Tips & Limitations

• BYOM (Bring Your Own Model): The skill does not include a model; you must provide your own VRM file to see the character.
• API Keys: To enable speech and lip-syncing, you must provide a valid ElevenLabs API key in your environment variables.
• Performance: Ensure your browser has hardware acceleration enabled to handle 3D rendering and animation blending smoothly.
• Licensing: Most included animations originate from Mixamo. Please adhere to their non-commercial usage terms when distributing your avatar setups.