Google’s Gemini app for Mac can now operate your computer for you
Google 2015 logo (Wikimedia Commons, public domain)
A new computer use feature lets Gemini work across Mac apps, local files and the web, starting with paid subscribers in the US
Google has given its Gemini app for macOS the ability to actually use your computer. The new feature, called computer use, lets the AI assistant carry out multi-step tasks that span local files, native desktop apps and the web.
According to Google, the feature can help users get multi-step tasks done without hopping between windows themselves. The company frames it as a way to take the tedious, repetitive parts of computer work off your plate.
Gemini can help you “get multi-step tasks done across your local files, desktop apps, and the web,” Google said.
How Gemini’s computer use works
Under the hood, the assistant performs the same actions a person would. It clicks, types, navigates menus and fills in forms. Crucially, it does this in the background rather than hijacking your screen.
Gemini runs its work in its own background window, and that window streams a live preview into your chat. So you can keep writing an email or editing a spreadsheet while the AI grinds through a chore in a separate lane.
Google suggested a handful of use cases. These include speeding up workflows, filling in forms, building presentations and listing items for sale online.
Google also pointed to tasks like organizing files and batch processing forms.
The feature launched on November 19, 2026. It is rolling out now in English for users in the US.
There is a catch on access. Using computer use requires a paid subscription to Google AI Plus, Pro or Ultra, so free-tier users will be watching from the sidelines for now.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
Control is the selling point
Users decide which apps Gemini is allowed to touch. Permissions are granted app by app, so you can let the assistant into your browser and presentation software while keeping it out of, say, your banking app.
The live preview doubles as a safety feature. Because you can follow along in real time, you are not left guessing what the AI did after the fact.
Users can also pause or stop an action midway. If Gemini starts heading somewhere you did not intend, you can pull the plug before it finishes.
That combination of scoped permissions, visibility and a kill switch is the core of Google’s pitch. The company is emphasizing security and user control as much as raw capability.
Late to a crowded party
Gemini did not start out with this power on the Mac. Google first released the Gemini macOS app in April 2026, and it already shipped with some basic automation features under the Gemini Tasks banner.
Google layered on more in June 2026 with its Spark and Tasks features. Computer use is the next step, moving the assistant from helping inside a chat window to acting across the whole desktop.
Google is not first here. OpenAI and Anthropic both introduced similar desktop automation tools before Google, which makes this a catch-up move rather than a land grab.
What this means for users and the AI race
The paywall is also worth watching. By tying computer use to Google AI Plus, Pro and Ultra, Google is using one of its most practical agent features as a reason to upgrade.
The open questions are practical ones. How well does Gemini handle messy, real-world apps that were never designed with AI agents in mind? How quickly will Google expand access beyond English-speaking US users?
For now, the rollout is limited and paywalled. But the direction is unmistakable: Google wants Gemini to be the assistant that does the work, not just the one that explains it.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.

