How to Set Up Browser Automation with OpenClaw
Browser automation unlocks powerful capabilities: web scraping, form filling, screenshot capture, and end-to-end testing. This guide shows you how to set up Playwright-powered browser automation in OpenClaw with proper error handling, stealth configuration, and anti-bot measures.
Why This Is Hard to Do Yourself
These are the common pitfalls that trip people up.
Headless browser setup
Chromium dependencies, sandboxing flags, and font rendering all need correct configuration
Flaky selectors
Websites change their DOM frequently. Hard-coded selectors break without warning.
Timeout and retry logic
Pages load at different speeds. Without proper waits, automations fail randomly.
Anti-bot detection
Many sites detect headless browsers and block automation. Stealth configuration is needed.
Step-by-Step Guide
Install Playwright in your OpenClaw environment
# In your OpenClaw directory:
npm install playwright
npx playwright install chromium
# For Docker, add to Dockerfile:
# RUN npx playwright install-deps chromiumCreate a browser automation skill
# ~/.openclaw/skills/browser-auto/skill.md
---
name: browser-auto
version: 1.0.0
description: Automates browser tasks using Playwright
author: your-name
permissions:
- network:outbound
- process:spawn
triggers:
- command: /browse
- pattern: "scrape|screenshot|fill out form"
---
## Instructions
You have access to a headless browser via Playwright.
When asked to perform browser tasks:
1. Launch browser with stealth settings
2. Navigate to the target URL
3. Perform the requested action
4. Return results (screenshots, extracted data, etc.)
5. Always close the browser when doneCreate the Playwright helper script
// ~/.openclaw/skills/browser-auto/scripts/browser.js
import { chromium } from 'playwright';
export async function browse(url, action) {
const browser = await chromium.launch({
headless: true,
args: ['--no-sandbox', '--disable-setuid-sandbox']
});
const page = await browser.newPage();
try {
await page.goto(url, { waitUntil: 'networkidle' });
if (action === 'screenshot') {
return await page.screenshot({ fullPage: true });
}
if (action === 'text') {
return await page.textContent('body');
}
// Add more actions as needed
} finally {
await browser.close();
}
}Configure browser settings for reliability
// Recommended Playwright settings:
const context = await browser.newContext({
viewport: { width: 1280, height: 720 },
userAgent: 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7)...',
locale: 'en-US',
timezoneId: 'America/New_York',
// Block unnecessary resources for speed:
// images, fonts, stylesheets
});Warning: Setting a custom userAgent helps avoid bot detection but does not make you invisible. Always respect robots.txt and website terms of service.
Add error handling and retries
// Wrap all browser actions in retry logic:
export async function browseWithRetry(url, action, maxRetries = 3) {
for (let attempt = 1; attempt <= maxRetries; attempt++) {
try {
return await browse(url, action);
} catch (error) {
if (attempt === maxRetries) throw error;
// Wait before retry (exponential backoff)
await new Promise(r => setTimeout(r, 1000 * attempt));
}
}
}Test your browser automation
# In OpenClaw chat:
/browse https://example.com screenshot
# Check for errors:
tail -f ~/.openclaw/logs/skills.log | grep browser-autoBrowser Automation Is Fragile
Headless browser config, anti-bot detection, flaky selectors, and retry logic — browser automation has many failure modes. Our experts build robust automations that actually work in production.
Get matched with a specialist who can help.
Sign Up for Expert Help →