Livekit Voice

LiveKit real-time voice and video infrastructure — create rooms, generate JWT access tokens, manage participants, and record sessions. Open source WebRTC for...

Security Scan

The skill implements expected LiveKit operations, but there are coherence issues (missing declared LIVEKIT_URL requirement) and it requires a sensitive API secret that grants broad room-admin capabilities—review before installing.

Purpose & Capability

Name, description, and code all match a LiveKit integration (create rooms, tokens, egress). However the registry metadata omits LIVEKIT_URL even though the script and SKILL.md require it; this mismatch is inconsistent and will break runtime behavior if not provided.

Instruction Scope

SKILL.md instructs only running the included Python CLI, and the CLI's actions are limited to LiveKit API calls (room management, token generation, listing participants, starting recordings). There is no code that reads unrelated files or exfiltrates data to third-party endpoints beyond the configured LiveKit URL.

Install Mechanism

No install spec — instruction-only with a single Python script. Nothing is downloaded from external URLs or installed automatically, which minimizes supply-chain risk. The included script will run with Python on the agent host.

Credentials

The script legitimately needs LIVEKIT_API_KEY, LIVEKIT_API_SECRET, and LIVEKIT_URL, but the registry metadata only lists the first two — an omission. LIVEKIT_API_SECRET is highly privileged (used to sign JWTs and the script generates tokens with roomAdmin/roomRecord grants for API calls). Ensure you trust the skill owner and the target LiveKit server before providing secrets.

Persistence & Privilege

always is false and the skill doesn't attempt to modify other skills or system-wide settings. It requires no persistent installation or elevated platform privileges.

What to consider before installing

This skill appears to do what it says (manage LiveKit rooms and tokens), but take these precautions before installing: - Fix the metadata gap: the script requires LIVEKIT_URL but the registry metadata omitted it; ensure you set LIVEKIT_URL to a trusted LiveKit endpoint. - Treat LIVEKIT_API_SECRET as highly sensitive: it can be used to generate management-level tokens. Only provide it if you trust the skill author and host environment. - Prefer using least-privilege credentials or a dedicated service account for agent usage rather than an owner/admin secret. - Review the included script locally to verify behavior (it only calls the configured LiveKit server). Run the skill in a sandboxed environment first. - If you will record to cloud storage (s3://...), double-check output paths and access controls so recordings don't leak to unintended buckets.

LiveKit real-time voice/video infrastructure for OpenClaw agents. Create rooms, generate tokens, manage participants, and integrate with voice AI platforms.

What is LiveKit?

LiveKit is an open-source WebRTC infrastructure platform for building real-time audio/video applications. It powers voice AI agents, video conferencing, live streaming, and more.

Self-hosted vs Cloud:

• LiveKit Cloud — Managed service, no infrastructure to maintain
• Self-hosted — Deploy on your own servers via Docker/Kubernetes

Requirements

Variable	Required	Description
LIVEKIT_API_KEY	✅	LiveKit API key
LIVEKIT_API_SECRET	✅	LiveKit API secret
LIVEKIT_URL	✅	LiveKit server URL (e.g. wss://your-project.livekit.cloud)

Quick Start

# Create a room
python3 {baseDir}/scripts/livekit_api.py create-room my-room

# Create room with options
python3 {baseDir}/scripts/livekit_api.py create-room my-room --max-participants 10 --empty-timeout 300

# Generate access token for a participant
python3 {baseDir}/scripts/livekit_api.py token my-room --identity user123 --name "John"

# Generate token with specific grants
python3 {baseDir}/scripts/livekit_api.py token my-room --identity agent --can-publish --can-subscribe

# List active rooms
python3 {baseDir}/scripts/livekit_api.py list-rooms

# List participants in a room
python3 {baseDir}/scripts/livekit_api.py participants my-room

# Delete a room
python3 {baseDir}/scripts/livekit_api.py delete-room my-room

# Start recording (Egress)
python3 {baseDir}/scripts/livekit_api.py record my-room --output s3://bucket/recording.mp4

Commands

create-room <name>

Create a new LiveKit room.

• --max-participants N — limit participants
• --empty-timeout N — seconds before empty room auto-closes (default 300)

token <room>

Generate a JWT access token for a participant.

• --identity ID — participant identity (required)
• --name NAME — display name
• --can-publish — allow publishing audio/video
• --can-subscribe — allow subscribing to others
• --ttl N — token TTL in seconds (default 3600)

list-rooms

List all active rooms with participant counts.

participants <room>

List participants in a room with their connection state and tracks.

delete-room <name>

Delete/close a room and disconnect all participants.

record <room>

Start an Egress recording of a room.

• --output URL — output destination (S3, GCS, or local path)

Voice AI Integration

LiveKit is the backbone for many voice AI platforms:

• Vapi — Uses LiveKit for real-time voice AI agent calls
• ElevenLabs — Stream TTS audio into LiveKit rooms
• OpenAI Realtime — Connect GPT-4o voice to LiveKit participants

Agent Pattern

1. Create a LiveKit room
2. Generate tokens for both human and AI agent
3. AI agent joins, subscribes to human audio
4. Process audio → STT → LLM → TTS → publish back
5. Result: real-time voice conversation with AI

Credits

Built by M. Abidi | agxntsix.ai YouTube | GitHub Part of the AgxntSix Skill Suite for OpenClaw agents.

📅 Need help setting up OpenClaw for your business? Book a free consultation