Skip to main content
It can perform actions like navigating to different pages, clicking on buttons, and filling out forms to test your applications. To start using the Agent-controlled Browser, ask Kombai to perform a task in the browser.

​
Capabilities

Kombai can identify interactive elements in your interface and test clickabiltity of these elements. It can perform clicks, double clicks, right-clicks and hover actions on any visible element on the webpage like buttons, links, and form fields.
Kombai can identify an input field and type text into it. It fills out forms, submits data, and interacts with text areas and search boxes.
Kombai can scroll lengthy pages to reveal more content, find specific elements, and explore large documents.
Kombai can take screenshots to understand page layouts, verify visual elements, and provide visual confirmation of its browser actions.
Kombai can monitor the browser console to troubleshoot problems and check page behavior by reading JavaScript errors, network warnings, and debugging output.
Kombai can monitor the page’s HTTP requests and responses, allowing it to track API calls, analyze payloads, check status codes, and diagnose network issues.
Kombai can track the page’s performance metrics, including load times, resource usage, and user experience indicators, to help identify and optimize performance bottlenecks.

​
How to switch to Agent-controlled browser

  1. Click the
    icon in the chat input box.
  2. Click Agent-controlled browser.

​
FAQ

​
I am unable to select the DOM element in the agent-controlled browser

The agent-controlled browser is entirely controlled by Kombai to execute tasks. If you need to manually select DOM elements, please use the user-controlled browser.