Google announced on November 19, 2026, that it was beginning an English-language U.S. rollout of computer use in the Gemini app for Mac. The feature is designed to carry out multi-step tasks across local files, desktop apps and the web by clicking, typing, scrolling, navigating menus and filling forms in apps a user permits. Google’s announcement
What Gemini computer use can do on a Mac
Gemini can act across permitted apps rather than merely give instructions in a chat: Google describes workflows that move information between apps and use local files or websites. Its examples include extracting shipping details from PDFs, gathering dashboard figures for a presentation and preparing photo edits.
Google’s promotional demonstration follows a hand-drawn butterfly sketch through modeling and rigging in Blender, an augmented-reality web preview and a printable ring file. That sequence illustrates the workflow Google chose to show.
Mac, account and subscription requirements
The Gemini app for Mac requires an Apple silicon Mac running macOS Sequoia 15.0 or later, with at least 8 GB of RAM and 200 MB of available disk space, according to Google’s Mac app requirements. Those are app requirements; they do not guarantee how well a particular computer-use task will perform.
The feature has separate eligibility rules. Google says users must be at least 18, use a personal Google Account, access Gemini in English in the United States, and subscribe to Google AI Plus, Google AI Pro or Google AI Ultra. Work and school accounts are not eligible for the feature at present, according to Google’s feature help.
macOS must also have Accessibility and Screen & System Audio Recording permissions enabled for the feature if they are not already on.
App permissions, controls and risks
On a first request to use an app, Gemini offers three choices: deny access, allow it once or always allow it. An always-allowed app stays on the permitted list until the user removes it in settings. Gemini works in a background or virtual screen with a live preview, and users can stop a task; Google says it requests confirmation for sensitive actions.
Give Gemini access only to the apps a task needs, and watch the live preview: Google warns that on-screen content can influence actions across permitted apps.
Google also warns that Gemini can make mistakes, encounter actions that cannot be reversed, and may not be able to back up every file. When temporary backups are available, they are kept for 24 hours or until the next prompt. The agent may encounter logged-in sessions and visible multi-factor authentication codes, while untrusted content on screen in one app could manipulate it into acting in another permitted app. These risks make app permissions consequential even when a task has a live preview and confirmation prompts.