avatar

What This Skill Does

The Avatar skill provides an advanced, multi-modal interface for the OpenClaw AI agent. By integrating Simli for real-time, lip-synced video generation and ElevenLabs for high-quality text-to-speech synthesis, it transforms standard text-based interactions into a natural, face-to-face conversation. Beyond the visual and auditory experience, the skill processes information in a dual-format structure, allowing the avatar to deliver a concise, conversational summary while simultaneously rendering a comprehensive markdown document in a dedicated side-panel. This ensures that users receive both the human-like interaction of an assistant and the structured, persistent data they need for professional workflows.

Installation

To integrate the Avatar skill into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/johannes-berggren/avatar

Once installed, you must configure your environment variables to authenticate with the underlying service providers. Ensure you have active API keys for both Simli and ElevenLabs. Add the following to your shell profile or environment configuration:

export SIMLI_API_KEY="your_simli_key"
export ELEVENLABS_API_KEY="your_elevenlabs_key"

After configuring, launch the service using openclaw-avatar and navigate to http://localhost:5173 in your web browser to start your first session.

Use Cases

This skill is ideal for professional environments where clarity and personal connection are paramount. Common use cases include:

• Executive Briefings: Start your day with a visual summary of your calendar, emails, and pending project tasks.
• Customer Support & FAQs: Provide a friendly, conversational interface for users to query technical documentation or company policies.
• Interactive Training: Use the avatar to deliver structured lesson plans that pair spoken explanations with dynamic visual markdown guides.
• Accessibility: Provide an alternative to text-only interfaces by offering a robust auditory and visual output stream.

Example Prompts

1. "Summarize the latest project updates from the engineering team and outline our deadlines for this week."
2. "Can you explain the current status of my outstanding customer support tickets and highlight any that are overdue?"
3. "Give me a quick breakdown of my inbox and suggest which emails require immediate attention today."

Tips & Limitations

• Response Format: Always ensure your agent logic splits output into <spoken> and <detail> tags. Failure to do so will result in poor user experience, as the spoken audio will include unwanted markdown formatting.
• Networking: This skill relies on external streaming services; ensure you have a stable high-speed internet connection for real-time lip-syncing performance.
• Resource Management: Video rendering can be resource-intensive. If experiencing lag, ensure your local development server is not competing for GPU cycles with other heavy processes.