Browser Automation Tutorial

In this guide we cover OpenClaw's browser automation—navigating sites, filling forms, extracting data, and building web workflows. Plan for about 25–30 minutes.

Here's what you'll know how to do when you're done:

• Navigate websites - Automatically browse and interact with web pages
• Fill forms - Automate form completion and submission
• Extract data - Scrape and parse information from websites
• Use Canvas - Leverage the visual Canvas workspace for complex interactions
• Build automation workflows - Create reusable web automation patterns

Before starting:

• OpenClaw installed and running - Complete Getting Started Tutorial
• Browser control enabled - Browser automation requires Chrome/Chromium
• Basic web knowledge - Understanding of HTML, CSS selectors helpful but not required

Browser control is typically enabled by default. Verify it's working:

openclaw agent --message "Navigate to https://example.com and tell me what you see"

If browser control isn't working, check the Troubleshooting Guide.

Browser Configuration

Configure browser settings in your OpenClaw config:

{
  "browser": {
    "enabled": true,
    "headless": false,
    "timeout": 30000
  }
}

Let's start with basic web navigation. Ask OpenClaw to visit a website:

Navigate to https://news.ycombinator.com and summarize the top 5 stories

OpenClaw will:

1. Open the browser
2. Navigate to the URL
3. Read the page content
4. Extract and summarize information

Interactive Navigation

You can guide OpenClaw through websites:

• "Click on the 'About' link"
• "Scroll down to find the pricing section"
• "Take a screenshot of the page"
• "Find all links on this page"

OpenClaw can automatically fill and submit forms. Here's how:

Simple Form Example

Ask OpenClaw to fill a form:

Go to https://example.com/contact and fill out the contact form with:
- Name: John Doe
- Email: john@example.com
- Message: Hello, I'm interested in your services

Complex Forms

For complex forms, provide detailed instructions:

• "Fill the registration form with my details"
• "Select 'Premium' plan from the dropdown"
• "Check the terms and conditions checkbox"
• "Submit the form"

OpenClaw will identify form fields and fill them appropriately.

Extract data from websites for analysis or storage:

Extract Text Content

Go to https://example.com/article and extract:
- The article title
- The author name
- The main content
- All links in the article

Extract Structured Data

Extract data into structured formats:

Go to https://example.com/products and extract all products with:
- Product name
- Price
- Description
- Image URL
Save this as a JSON file

Scrape Multiple Pages

Automate scraping across multiple pages:

Visit https://example.com/products and:
1. Extract all product links from page 1
2. Visit each product page
3. Extract product details
4. Save all data to products.json

Canvas provides a visual workspace for complex browser interactions:

What is Canvas?

Canvas is OpenClaw's visual interface that allows the agent to:

• Render interactive UIs
• Display visual information
• Create custom interfaces
• Show browser screenshots

Access Canvas

Canvas is accessible via:

• Control UI at http://localhost:18789
• macOS menu bar app
• iOS/Android companion apps

Canvas with Browser

When using browser automation, Canvas can:

• Show browser screenshots
• Display extracted data visually
• Render custom visualizations
• Provide interactive feedback

Price Monitoring

Monitor prices on e-commerce sites:

Every day at 9am:
1. Visit https://example.com/product/123
2. Extract the current price
3. Compare with previous price
4. Send me a notification if price dropped

News Aggregation

Aggregate news from multiple sources:

Visit these news sites:
- https://news.ycombinator.com
- https://www.reddit.com/r/programming
Extract top 5 stories from each
Create a summary document

Research Automation

Automate research tasks:

Research "OpenClaw AI assistant" and:
1. Search Google for recent articles
2. Visit top 5 results
3. Extract key information
4. Create a research summary document

Respect Website Terms

• Always respect robots.txt
• Don't overload servers with requests
• Add delays between requests when scraping
• Follow website terms of service

Error Handling

• Handle page load failures gracefully
• Verify elements exist before interacting
• Use timeouts for slow-loading pages
• Log errors for debugging

Performance

• Use headless mode for faster execution
• Cache frequently accessed data
• Optimize selectors for faster element finding
• Close browser tabs when done

Browser Not Starting

• Check Chrome/Chromium is installed
• Verify browser permissions
• Check Gateway logs: openclaw gateway logs
• See Troubleshooting Guide

Elements Not Found

• Wait for page to fully load
• Check element selectors are correct
• Use browser dev tools to verify selectors
• Try more specific selectors

Slow Performance

• Enable headless mode
• Reduce screenshot frequency
• Optimize page interactions
• Use faster selectors

Now that you understand browser automation, explore these related topics: