Gemini for Mac computer use, Digimart Global (United States): Google has introduced a feature that allows the Gemini app for Mac to operate other programs on its own. Google announced the feature, called computer use, on November 19, according to a publisher report that has not been independently verified.
The feature lets Gemini open apps and websites, then click and type through repetitive tasks. To start a task, users add “@Computer use” to their prompt.
Google says Gemini works inside an isolated background window with a live preview in the chat, so users can continue other work while still seeing what the agent is doing.
Access, limits and approvals
Users choose which apps Gemini can access. Google says the agent can be paused or stopped at any time, and key actions can be approved before they happen.
According to 9to5Google, the agent will pause for logins and payments. The publisher report notes that Google’s announcement does not say whether the agent stores what it sees on screen or uses that content for AI training.
For existing tasks that work with connected folders, the Gemini help page warns users not to enter sign-in or payment details into the task thread. Temporary backup files created there are deleted after 24 hours at the latest.
Requirements and availability
The rollout is limited to English in the US for now. It requires a Google AI Plus, Pro or Ultra subscription, a personal Google account for users 18 and older, and an Apple silicon Mac running macOS Sequoia 15 or later.
Google has not said when other countries will receive the feature. According to Macworld, the rollout is gradual, and the app version it tested did not yet include the feature.
Users who give an AI agent access to a Mac should review permissions in System Settings. Apple has reportedly planned to restrict such rights for AI agents.














