Browser Relay: Control Chrome/Firefox from OpenClaw
Automate web browsing, form filling, screenshots, and data extraction by giving your agent direct browser control.
What You Will Get
After this setup, your OpenClaw agent will be able to open web pages, click buttons, fill forms, take screenshots, and extract data from websites through a browser relay. This gives your agent the ability to interact with any website as if it were a human user.
Browser automation is incredibly powerful for tasks like price monitoring, form submissions, data scraping, automated testing, and capturing visual snapshots of web content. Your agent can navigate multi-step workflows on websites that do not have APIs.
The browser relay runs Chrome or Firefox in a headless mode and exposes control to your agent through a secure connection. You can watch the browser in real time through a live view panel, making it easy to debug and monitor what your agent is doing on the web.
Step-by-Step Setup
Connect a browser relay to give your agent web browsing capabilities.
Enable the Browser Relay Channel
In your RunTheAgent dashboard, navigate to Channels and select Browser Relay. This activates a managed browser instance that your agent can control. Choose between Chrome and Firefox based on your compatibility needs. Chrome is recommended for most use cases.
Configure Browser Settings
Set the browser viewport size, user agent string, and default language. Configure proxy settings if you need to route traffic through a specific location. Enable or disable JavaScript, cookies, and image loading based on your use case. These settings affect how websites see and respond to your agent.
Set Up Navigation Commands
Define navigation templates for common tasks. Each template specifies a URL pattern and a sequence of actions like clicking, typing, or waiting. For example, create a template that navigates to a search engine, enters a query, and extracts results. Templates make it easy to reuse browser workflows.
Configure Screenshot Capture
Enable screenshot capture so your agent can take visual snapshots of web pages. Set the default image format (PNG or JPEG), quality, and whether to capture the full page or just the visible viewport. Screenshots can be sent to users through any connected messaging channel.
Set Up Data Extraction
Configure CSS selectors or XPath expressions to extract specific data from web pages. Define extraction templates that pull text content, link URLs, image sources, or structured data from any page. Your agent can use these templates to gather information from websites on demand.
Configure Security Boundaries
Set URL allowlists and blocklists to control which websites your agent can visit. Block known malicious domains and restrict the browser to specific domains if your use case is narrow. Enable content security policies to prevent the browser from executing potentially harmful scripts.
Test a Browser Task
Ask your agent to visit a website and extract some information. Monitor the live view panel to watch the browser in action. Verify that screenshots are captured correctly and data extraction returns the expected results. Check the RunTheAgent logs for any navigation errors or timeouts.
Tips and Best Practices
Use Waits Instead of Fixed Delays
Configure the browser to wait for specific elements to appear rather than using fixed time delays. This makes automation more reliable and faster since it does not wait longer than necessary.
Handle CAPTCHAs Gracefully
Some websites present CAPTCHAs. Configure your agent to detect CAPTCHA pages and either notify you for manual intervention or skip the task rather than getting stuck in a loop.
Rotate User Agents
If you are browsing multiple sites frequently, rotate user agent strings to avoid being flagged as a bot. The browser relay supports automatic user agent rotation from a predefined list.
Frequently Asked Questions
Related Pages
Ready to get started?
Deploy your own OpenClaw instance in under 60 seconds. No VPS, no Docker, no SSH. Just your personal AI assistant, ready to work.
Starting at $24.50/mo. Everything included. 3-day money-back guarantee.
