Give Your AI Agent Eyes
Turn any phone, laptop, or tablet camera into a visual sensor for OpenClaw agents. Snapshot, stream, and perceive — all peer-to-peer, no cloud required.
How It Works
Add the Skill
Give your agent the ClawCam skill URL to enable camera access.
Setup & Activate
Install the CLI, set up a camera device, and scan the QR code on your phone.
Agent Gains Vision
Your agent captures frames and analyzes the physical world in real time.
Built for Agent Vision
Workspace Perception
Mount a phone on your desk and let agents observe, guide, and assist with physical tasks.
Environment Sensing
Give agents spatial awareness — monitor rooms, detect changes, and understand surroundings.
Remote Visual Interaction
Point a camera at equipment or sites for remote agent-powered inspection and guidance.
Quick Start
Add skill to your agent
https://clawcam.pages.dev/skill.md
Provide this URL to your AI agent to enable camera capabilities.
Install the CLI
npm install -g clawcam
Set up a camera device
clawcam setup my-camera
Then open the generated URL on your phone and grant camera permission.
Try It Now
Scan this QR code with your phone to instantly activate a camera. Once connected, you'll be redirected to the live viewer.
Key Capabilities
Peer-to-Peer
Direct WebRTC connection. No video passes through any server.
No Cloud Required
Everything runs locally. Your video data stays on your network.
Instant Snapshots
Capture single frames via DataChannel for fast vision model analysis.
Multi-Device
Connect multiple cameras simultaneously as different visual inputs.
Cross-Platform
Any device with a browser and camera works — phones, laptops, tablets.
Low Latency
Sub-second video delivery for real-time agent perception and response.