OpenClaw includes powerful browser automation capabilities and a visual Canvas workspace. These features allow the agent to interact with websites, automate tasks, and provide a visual interface for complex operations.
The browser tool enables Moltbot to:
- Navigate Websites - Visit and interact with any website
- Fill Forms - Automatically fill and submit web forms
- Extract Data - Scrape and extract information from pages
- Click Elements - Interact with buttons, links, and UI elements
- Take Screenshots - Capture page screenshots
- Execute JavaScript - Run custom JavaScript on pages
Enabling Browser Control
Enable browser control in your configuration:
{
"browser": {
"enabled": true,
"color": "#FF4500"
}
}
Use Cases
- Automate form submissions
- Monitor websites for changes
- Extract data from web pages
- Perform web-based research
- Automate repetitive web tasks
Canvas is Moltbot's visual workspace feature:
- Visual Interface - Agent-driven visual workspace
- Interactive Elements - Buttons, forms, and controls
- Real-Time Updates - Dynamic content updates
- Node Integration - Accessible via iOS/Android nodes
Canvas Host
The Canvas host serves visual interfaces:
- HTTP file server on port 18793 (default)
- Serves
/__openclaw__/canvas/ paths
- Accessible via Control UI or nodes
- Secure by default (loopback or authenticated)
Form Automation
Ask Moltbot to fill and submit forms:
- "Fill out the contact form on example.com"
- "Submit this application form with my details"
- "Book a flight for next week"
Data Extraction
Extract information from websites:
- "Get the latest prices from this product page"
- "Extract all email addresses from this page"
- "Scrape the table data from this website"
Web Monitoring
Monitor websites for changes:
- Check for product availability
- Monitor price changes
- Track website updates
Browser control is powerful and should be used carefully:
- Sandboxed by Default - Browser runs in isolated environment
- Session Isolation - Each session has its own browser context
- Tool Policies - Control browser access per session
- Trusted Sessions Only - Browser typically restricted to main session
For untrusted sessions (groups), browser access is typically disabled by default.
Visual Workspace
Canvas provides a visual interface where:
- Agents can render interactive UIs
- Users can interact with visual elements
- Real-time updates reflect agent actions
- Complex workflows become visual
Node Integration
Canvas is accessible via:
- iOS Node - View Canvas on iPhone/iPad
- Android Node - Access Canvas on Android devices
- Control UI - Browser-based Canvas access
Common browser issues and solutions:
- Linux Issues - May require additional dependencies (see Linux browser troubleshooting documentation)
- Permissions - Ensure browser has necessary permissions
- Headless Mode - Browser runs headless by default
- Timeout Issues - Adjust timeout settings if pages load slowly