Overview
- Google added 'computer use' to the Gemini Mac app on Thursday, Nov. 19, 2026, and the capability is rolling out in U.S. English for Google AI Plus, Pro, and Ultra subscribers.
- The feature lets Gemini simulate clicks and typing to navigate menus, fill forms, move files, and complete workflows across local desktop apps and web pages.
- Actions run inside an isolated background window with a live preview so your main screen and cursor stay free while you watch the agent work.
- Users can choose which apps Gemini can access and can pause, stop, or approve steps; Google says it will automatically pause for logins and payments and require confirmation for sensitive actions.
- The release follows a broader Nov. 17 rollout of free Gemini Tasks in six countries and joins existing desktop agents from other AI companies, which could speed routine work but raise questions about platform reach and data handling.