Browser control by the Copilot
The AI Copilot can carry out tasks in your browser for you: open a page, click something, fill in a form, read content or take a screenshot. The Companion runs these steps in your Chrome, using your own sign-in on the sites involved.
What the Copilot can do
Open pages, go back and forward · click, type, press keys, select, scroll, drag · read and extract content · wait for elements · open, switch and close tabs · handle dialogs and cookie banners · download and upload files · take screenshots.
You stay in control
- Its own tab. The Copilot works in a new tab in your current window, not in the tab you are using.
- Visible notice. While the Copilot controls a tab, the tab shows aYOUne steuert diesen Tab (aYOUne is controlling this tab) with a Stopp button. One click ends the control immediately.
- Automatic end. If no new command arrives for two minutes, control ends by itself.
- Approvals. Every command is checked before it runs and assigned to one of three levels: read, interact or submit. Anything not explicitly allowed is refused. Submit covers, for example, clicking a submit button, typing into password or payment fields and uploading files.
- Asking you first. If a step needs your consent, you receive a notification with approve and reject buttons. Without your approval the Copilot does not run the step and does not retry it.
Chrome's notice
For clicks and typing that a web page should treat as real user input, the Companion briefly attaches to the tab. Chrome then shows a notice at the top that the extension "started debugging this browser". This is expected and disappears after the step.
Naming your browser
If you use the Companion on several computers, give each browser its own name under Settings → Account & tenant → Name of this browser (remote control). aYOUne then shows which browser is carrying out a task.