Browser Agent (CDP) — OpenClaw Plugin
AI-powered browser automation for OpenClaw via Chrome DevTools Protocol.
What it does
Gives your OpenClaw agent the ability to control a real browser:
- Navigate to URLs
- Click elements, fill forms, select dropdowns
- Take screenshots
- Extract text, HTML, and page data
- Execute JavaScript
- Batch multiple commands
Powered by agent-browser — a fast native Rust CLI.
Install
openclaw plugins install @openclaw-plugins/browser-agent
Setup
-
Install agent-browser:
npm install -g agent-browser agent-browser install -
To control an existing Chrome, launch it with CDP enabled:
chrome --remote-debugging-port=9222 -
Connect:
agent-browser connect 9222
Tools Provided
| Tool | Description |
|---|---|
browser_open | Navigate to a URL |
browser_snapshot | Get page accessibility tree with refs |
browser_click | Click element by ref |
browser_fill | Fill form field |
browser_screenshot | Capture screenshot |
browser_get | Extract page data |
browser_eval | Run JavaScript (optional) |
browser_batch | Multi-command execution |