Gemini can now use your Mac: Google ships computer use on macOS
Google added computer use to Gemini for Mac on Thursday: native app control, file-level tasks and live previews, English-only in the U.S. and gated behind AI Plus, Pro or Ultra plans.
By Amara Osei
3 min read
What's News
- Google announced on Thursday that Gemini's macOS desktop app now offers computer use, acting inside native Mac apps and browsing websites in the background.
- The feature launches in English in the U.S. and requires the Gemini app for Mac plus a Google AI Plus, Pro or Ultra subscription.
- Google began the broader rollout of Gemini's "Tasks" feature this week, which signs users into websites, makes online purchases and books appointments.
- Google announced its revamped "CC" agent in September and only recently began admitting early testers.
- OpenAI, Anthropic and Meta have already offered agentic features in their desktop applications for some time.
Google on Thursday brought computer use to Gemini's desktop app on macOS, letting the model finish jobs inside native Mac apps and browse websites on a user's behalf. The feature handles multi-step work across local files, applications and the web, according to Google.
Access is narrow at launch. The rollout starts in English, in the U.S., and requires both the Gemini app for Mac and a Google AI Plus, Pro or Ultra subscription. Google didn't say when the feature would expand to other markets or platforms.
Google is also late to the category. OpenAI, Anthropic and Meta have already been offering agentic features in their desktop applications for some time, and this week Google began the broader rollout of two more agentic products of its own.
What can Gemini actually do on a Mac?
Users give Gemini an instruction and watch it work. The action shows up in a live preview inside the chat window, so the screen and cursor stay free for other work while the agent runs in the background.
Control stays with the person at the keyboard. As with other agents offering computer use, users can pause, stop and approve Gemini's actions before they happen.
Computer use differs from those two products in where the action happens. It lets an agent take actions directly on the user's computer, which in Gemini's case means a Mac, spanning local files, installed apps and the open web.
Which tasks does Google expect people to hand off?
Google points to tedious work and repeatable workflows that span files and apps. Its examples:
- Filling out forms.
- Repeating the same task across a batch of files.
- Organizing files into folders based on specific criteria.
- Creating slideshows from files or information in online dashboards.
Each of those runs across the user's own documents and installed software rather than inside a single browser tab.
What else did Google ship this week?
The company also began the broader rollout of Gemini's "Tasks" feature, the in-app tool that lets Gemini work more as an agent. Google says Tasks can help users sign into websites, make online purchases and book appointments.
Google announced its revamped "CC" agent in September, and only recently started admitting early testers. The company designed "CC" to help families stay organized across email, calendars, chats and tasks.
So the Mac release lands as the third agentic push in a single week for Google, after two features aimed at shopping, scheduling and household logistics.
Why does the subscription gate matter?
Computer use arrives as a paid feature, not a free one. Anyone who wants Gemini operating a Mac needs the Gemini app for Mac plus an AI Plus, Pro or Ultra plan, which limits the pool of users Google can measure the feature with on day one.
For companies watching agentic desktop tools, the gap Google is closing is specific: native app control, file-level operations and a live preview of what the agent is doing. What Google isn't closing yet is reach, since the feature stops at the U.S. border and at the paywall.
What comes next?
The unanswered questions are timing and platform. Google has committed to neither a wider rollout nor a date, so adoption beyond U.S. English speakers depends on ports to other operating systems and the spread of subscription plans.
Google's own roadmap this week favors shipping agents into the desktop layer, where OpenAI, Anthropic and Meta got there first. The test now is whether AI Plus, Pro and Ultra subscribers actually turn their Macs over, and whether Google lowers the wall around computer use before rivals widen theirs.
Source: TechCrunch
More from Amara Osei
Show full bio
Senior reporter covering companies, deals and strategy in entrepreneurship, startups and venture funding at Business Bearings.
2,422 articles