Multiple tasks and browsers in UI agents
- Updated: 2026/02/25
The UI Agents can operate multiple tasks by using multiple Run actions, and multiple browsers can be used for operations across URLs.
Chaining multiple tasks and persisting browser state
After breaking down a complex goal, you can add multiple Run actions in your UI Agent
and specify the task each of them needs to perform in the
Goal. If these tasks need to be performed on the same
browser, specify the Starting URL for the first
Run action and leave it empty for later actions.
Otherwise, type a Starting URLto have the agent open that URL
before running the task.
Example: multiple tasks on the current webpage of a browser session.


Example: multiple tasks on different webpages within the same browser session.


Working with multiple browsers
With the UI Agents, you can open multiple browser sessions and have UI
agents operate across each of them (sequentially). This is useful when the workflow
requires data to be referenced between two or more webpages without saving them in
an intermediate file.

