Manage real-time voice and video infrastructure for AI agents using LiveKit WebRTC.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install livekit-voice
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install livekit-voice using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is LiveKit Voice?
LiveKit Voice is an essential addition to Openclaw Skills, providing robust WebRTC infrastructure for building real-time audio and video applications. It allows developers to create rooms, manage participant access via JWT tokens, and integrate with leading voice AI platforms like Vapi and ElevenLabs.
Whether using LiveKit Cloud or a self-hosted instance, this skill simplifies the complex task of handling real-time communication. It enables Openclaw Skills to function as the backbone for sophisticated voice-enabled AI agents, providing low-latency interactions and professional-grade session management.
LiveKit Voice Use Cases
- Building real-time voice AI agents that interact with humans via WebRTC.
- Automating the creation and teardown of video conferencing rooms.
- Generating secure access tokens for participants in a live session.
- Recording AI-human conversations for training or compliance using LiveKit Egress.
- Integrating OpenAI Realtime or ElevenLabs TTS into a live communication stream.
How LiveKit Voice Works
- The user provides LiveKit credentials to the Openclaw Skills environment.
- The skill initializes a connection to the LiveKit server (Cloud or self-hosted).
- A new room is created with specific participant limits and timeouts via the CLI.
- JWT access tokens are generated with granular permissions (publish/subscribe) for agents and users.
- Participants join the room, and the agent begins processing audio/video tracks in real-time.
- Optional recording (Egress) is triggered to save the session to cloud storage.
LiveKit Voice Setup
To use this skill within Openclaw Skills, ensure you have your LiveKit credentials ready. Set the following environment variables:
export LIVEKIT_API_KEY="your_api_key"
export LIVEKIT_API_SECRET="your_api_secret"
export LIVEKIT_URL="wss://your-project.livekit.cloud"
Basic command to create a room:
python3 scripts/livekit_api.py create-room my-ai-room --max-participants 2
LiveKit Voice Data Schema & Taxonomy
The LiveKit Voice skill organizes communication metadata and session states as follows:
| Component | Description |
|---|---|
| Room Metadata | Includes name, participant count, and active status. |
| JWT Tokens | Secure strings containing identity, grants, and TTL. |
| Participant Data | Identity, connection state, and active media tracks. |
| Egress Records | S3/GCS paths or local paths for session recordings. |
LiveKit Voice Advanced Features
- Support for LiveKit Egress to automate session recordings to S3 or GCS.
- Granular JWT grant management for fine-tuned control over audio/video publishing.
- Seamless integration with Vapi and OpenAI Realtime for advanced Openclaw Skills workflows.
- Automatic room cleanup using the empty-timeout parameter to optimize resource usage.
- Multi-agent support where multiple AI participants can join a single room.
SKILL.md
Loading
Related Openclaw Skills
Linode/Akamai Management Skill
A comprehensive Python-based CLI tool for managing Linode (Akamai) cloud infrastructure, including compute, networking, and Kubernetes.
Lemon Squeezy
A comprehensive CLI tool for managing Lemon Squeezy digital products, orders, and subscriptions within Openclaw Skills.
Lead Gen Pipeline
An AI-powered automation skill for scoring leads and generating context-aware sales follow-ups.
Langfuse Observability
A professional v3 observability toolkit for OpenClaw agents featuring automatic tracing, cost analysis, and system health monitoring.
LLM Evaluator
An automated LLM-as-a-Judge system that scores Langfuse traces for quality, accuracy, and hallucination detection.
LLM Evaluator
An automated LLM-as-a-Judge evaluation system for scoring AI response quality, accuracy, and hallucination using Langfuse.
Featured*
