claw-browser

What This Skill Does

The claw-browser skill provides OpenClaw agents with a sophisticated, visual Chromium-based web interface. Unlike traditional scrapers or HTTP request libraries, this tool renders full JavaScript, executes dynamic front-end frameworks, and supports single-page application (SPA) navigation. By acting as a true browser, it bypasses many common bot-detection challenges and ensures that the agent sees the web as a human would, complete with screenshot capabilities and an optional graphical user interface (visible mode).

Installation

Installation requires a standard Python environment and the Playwright browser engine. Execute the following in your OpenClaw terminal:

1. Install via CLI: clawhub install openclaw/skills/skills/canbirlik/claw-browser
2. Set up the environment: Create a virtual environment (python3 -m venv venv) and activate it (source venv/bin/activate).
3. Install dependencies: pip install -r requirements.txt.
4. Initialize Browser Binaries: This is the most critical step. Run playwright install chromium. Without this command, the agent will lack the necessary browser engine to render web pages.

Use Cases

• Automated Research: Have your agent read complex, JS-heavy documentation or search results that standard crawlers cannot interpret.
• Visual Verification: Capture snapshots of website states to confirm UI changes or log evidence of activity.
• Interactive Demos: Run the browser in visible mode to allow users to watch the agent navigate, click, and interact with web elements in real-time.
• Dynamic Data Extraction: Extract information from sites that require authentication or interactive content loading.

Example Prompts

1. "Open https://www.python.org/ and summarize the latest news headlines for me using the claw-browser."
2. "Take a screenshot of the current pricing page on https://stripe.com and save it for my records."
3. "Open https://google.com in visible mode and search for 'OpenClaw documentation'."

Tips & Limitations

Always ensure the virtual environment is active before triggering browser actions. For WSL or Linux users encountering display errors, ensure your X11 server is configured correctly. Note that persistent high-frequency browsing may trigger rate limits or CAPTCHAs, so utilize this skill responsibly. Always rely on the --visible flag for troubleshooting UI-related issues or when demonstrating agent capabilities to stakeholders.