Skip to main content
Vibium is a brand-new browser automation framework for AI agents that is built on WebDriver BiDi, a W3C standard. As of v26.3.9, any agent can now use Vibium to connect to Kernel cloud browsers to navigate pages, fill forms, click buttons, and take screenshots. We’re really excited to keep working with Jason Huggins and to keep supporting open web standards.

Adding Kernel to existing Vibium implementations

If you already have a Vibium implementation, you can switch to using Kernel’s cloud browsers by creating a Kernel browser session and connecting Vibium to the returned webdriver_ws_url.

1. Install Kernel and Vibium

2. Create a Kernel browser

3. Connect Vibium to your Kernel browser

4. Use your Vibium automation

Use Vibium’s page methods with the Kernel-powered browser:

5. Clean up

When you’re done, close Vibium and delete the Kernel browser session:

Using Vibium CLI with Kernel

The CLI can connect directly to the Kernel browser’s WebDriver BiDi endpoint:

Using Vibium MCP server with Kernel

The MCP server reads the same environment variable, so AI agents can use a Kernel-powered browser:
Or in your Claude Desktop / Claude Code config:

Benefits of using Kernel with Vibium

  • No local browser management: Run automations without installing or maintaining browsers locally
  • Scalability: Launch multiple browser sessions in parallel
  • Stealth mode: Built-in anti-detection features for web scraping
  • Session state: Maintain browser state across runs via Profiles
  • Live view: Debug your automations with real-time browser viewing

Next steps