Skip to main content

Browser Automation

Control browser tabs, perform browser actions, and retrieve web page content.

Browser actions by design use CSS selectors to identify elements on a page. If you would like your application to operate on a higher Level of Autonomy, you can use the AI Prompt to dynamically identify HTML elements and their CSS selectors.

The following Prebuilt Workstation Specs include a virtual browser:

Browser Tabs​

Note: We recommend using the same browser profile for each tab to avoid issues with cookies and browser extensions.

Workstations can manage multiple browser tabs, allowing agents to perform tasks across multiple web applications.

  • List: List all tabs in the Workstation browser.
  • Open: Open a new tab in the Workstation browser.
  • Close All: Close all open tabs in the Workstation browser.
  • GoToURL: Navigate the active tab to go to a specific URL.
  • Retrieve: Retrieve a specific tab in the Workstation browser.
  • Close: Close a specific tab in the Workstation browser.
  • Active: Set a specific tab in the Workstation browser as the active tab.
  • URL: Navigate to a specific URL in the Workstation browser.
  • Refresh: Refresh a specific tab in the Workstation browser.

Browser Actions​

Note: Most browser actions require an active tab to be set.

Currently, Workstation browsers can complete the following actions:

  • Click: Click on an element in the active tab of the Workstation browser.
  • Scroll: Scroll to the location of the given query selector on the current active tab of the Workstation browser.
  • Hover: Hover over an element in the active tab of the Workstation browser.
  • Keyboard: Type text into an input field on the active tab of the Workstation browser.
  • Text: Input text into a field on the active tab of the Workstation browser.
  • Select: Select an option from a dropdown or multi-select element in the active tab of the Workstation browser.
  • HTML: Retrieve the HTML content of the active tab of the Workstation browser. A download link is returned in the response.
  • Screenshot: Screenshot the active tab of the Workstation browser. A download link is returned in the response.