TLDR
Google’s Gemini app for Mac can now type, click and complete tasks for users.
The feature runs in an isolated background window, so users can keep working.
Users pick which apps Gemini can access and can pause it at any time.
Gemini asks for approval before taking key actions.
Access is limited to US English users on AI Plus, Pro or Ultra plans with Apple Silicon Macs.
Google has added a computer use feature to its Gemini app for Mac. The new tool lets the AI assistant type, click and carry out repetitive tasks for users.
The feature runs inside an isolated background window. This means people can keep using their Mac for other work while Gemini handles a separate job.
The update was reported by Cult of Mac on Friday, November 20, 2026. It is the latest addition to the Gemini app that Google built for Mac computers.
What Gemini Can Now Do on a Mac
Google shared several examples of tasks the assistant can complete. These include color grading photos in a photo editing app and setting up complex 3D scenes.
Gemini can also fill out online forms and pull data out of PDF files. It can search through files stored on a user’s Mac to find relevant information.
Another example is building a presentation in Google Slides. The feature is aimed at jobs that usually require switching between several apps and websites.
Before this update, the Mac app already offered dictation. It could also use what was showing on a user’s screen as context when answering questions.
How Users Stay in Charge of the Assistant
Google said users remain in control throughout the process. People decide which apps Gemini is allowed to access before it begins working.
A live preview inside the chat window shows what the assistant is doing. Users can follow its progress as each task moves forward.
Users can also pause Gemini at any point during a task. The assistant asks for approval before it takes key actions.
View post by @GeminiApp on X
These approval requests give users a chance to stop or change a task before an important step. The controls work alongside the live preview and app access settings.
Similar tools are already offered by rival AI companies. Anthropic’s Claude and OpenAI’s ChatGPT both have computer use features that can operate apps on a computer.
With this release, Google now offers a comparable feature inside its own Mac app. Gemini can now act directly on a desktop in the same way as those competing assistants.
The feature is not open to all Gemini users yet. It is rolling out in English in the United States only.
Access is limited to people using personal Google accounts. They must also subscribe to the AI Plus, Pro or Ultra plan to use it.
There are hardware and software requirements as well. Computer use in Gemini needs an Apple Silicon Mac running macOS Sequoia 15 or later.

