Skip to main content
Your AI assistant can browse the web.

​
How it works

Every EasyClaw container includes a full browser that your AI assistant can control. This isn’t a simplified web scraper — it’s a real browser with JavaScript execution, cookie management, and visual rendering. When you ask your AI to “look up the weather” or “find the cheapest flight to London,” it opens the browser, navigates to the relevant pages, reads the content, and reports back.

​
Browser capabilities

The browser inside each container can:
  • Navigate — Go to URLs, follow links, use back/forward
  • Read — Extract text, tables, and structured data from pages
  • Interact — Click buttons, fill forms, scroll, select dropdowns
  • Wait — Handle dynamic content, JavaScript rendering, loading states
  • Screenshot — Capture visual snapshots of pages for analysis
Your AI reasons about what it sees on screen and takes appropriate actions, just like a human would.
The browser runs inside your private container. Cookies, sessions, and browsing history are isolated to your environment and not shared with anyone else.

​
Use cases

​
Research

Ask your AI to research a topic and it will browse multiple sources, compare information, and synthesize a summary.
“Research the best project management tools for a team of 10 and compare pricing”

​
Form filling

Your AI can fill out web forms on your behalf — booking appointments, submitting applications, or completing registrations.
“Go to the dentist website and book me an appointment for next Tuesday afternoon”

​
Monitoring

Set up your AI to periodically check web pages for changes — price drops, stock availability, news updates.
“Check if the PS5 is back in stock on Amazon and let me know”

​
Data extraction

Extract structured data from websites — product listings, contact information, schedules.
“Get me a list of all the speakers at the upcoming tech conference from their website”

​
Visual browsing

You can watch your AI browse in real-time from the EasyClaw web console. The browser’s display is streamed to your browser via a VNC connection, giving you a live view of what the AI is doing. This is useful for:
  • Debugging complex browsing tasks
  • Watching the AI fill out forms (to verify it’s doing the right thing)
  • Intervening if something goes wrong

​
Configuration

Browser behavior is controlled through ZeroClaw’s tool configuration: 
[tools.browser]
headless = true
viewport_width = 1920
viewport_height = 1080
timeout_seconds = 30
max_pages_per_session = 50
SettingDefaultDescription
headlesstrueRun without visible window (VNC still available)
viewport_width1920Browser window width
viewport_height1080Browser window height
timeout_seconds30Max time to wait for page loads
max_pages_per_session50Limit on pages visited per browsing session

​
Privacy

Browser sessions are completely isolated:
  • Each container has its own browser profile
  • No cookies or history are shared between users
  • Browsing data is wiped when your container is destroyed
  • Your AI’s browsing activity is not logged or monitored by EasyClaw