Gemini can now run your Mac apps while you keep working
Google's Gemini app for Mac can now click, type and fill in forms across apps and websites in its own window, starting with English-speaking US subscribers on paid plans.

Google has added computer use to its Gemini app for Mac, letting the AI click, type and fill in forms across desktop apps and websites on a user’s behalf, TechCrunch reported. It starts in the US, in English, for paid subscribers only.
Who gets it today, and what it needs
According to Google’s Gemini app release notes, dated November 19, 2026, the feature is rolling out in English in the United States for personal Google accounts held by users aged 18 and over. It requires a Google AI Plus, Pro or Ultra subscription and the Gemini app for Mac, which is downloaded from gemini.google/desktop. The release entry lists no other technical requirements, such as a minimum macOS version.

TechCrunch adds that Google did not say when the feature would reach other markets or platforms. Readers outside the US, on the free tier, or on Windows have no stated date.
A separate window, a live preview and a stop button
Google describes the agent as handling multi-step clicking, typing and form-filling across desktop apps and websites. It works in its own window, where the user can watch a live preview, stop it with one click, and keep using their own screen and cursor for other work.
The release notes also describe the controls. Users choose which apps Gemini may use, Gemini pauses when a login or a payment is needed, and it may ask for confirmation before sensitive actions. TechCrunch reports that users can pause, stop and approve actions before they happen.
The use cases Google suggests, per TechCrunch, are tedious ones: filling out forms, repeating a task across a batch of files, sorting files into folders by given criteria, and building slideshows from files or from information in online dashboards.
Google arrives after its rivals on the desktop
TechCrunch says Google is late with this launch, noting that OpenAI, Anthropic and Meta have offered agentic features in their desktop apps for some time. The Mac app itself is also fairly new. Google launched a native Gemini app for Mac on April 15, 2026, free for all users, according to TechCrunch’s coverage at the time. Reports from that launch, such as one from AlternativeTo, said that the first version could share windows and local files as context but did not seem able to take over the computer and act for the user. Seven months later, that gap is being closed.
The timing also fits a pattern in Google’s own releases this week. Gemini Tasks, the agent that works inside the app, was made free in six countries, and in the US it can call businesses and pay with Google Pay, as we reported on November 17. Tasks acts mostly through the web and phone calls, while the Mac feature reaches software that lives on the user’s own machine. The two cover different ground, and a paid subscription is the entry ticket for the Mac one, whereas Tasks was opened up for free.
One more detail from that week bears on trust. Google extended how long it keeps temporary chats and chats with history turned off from 72 hours to about two months, as our earlier article covered. An agent that can see a screen full of files and forms makes that retention policy more relevant. Neither the release notes nor TechCrunch says how screen content seen during a computer use session is stored or retained.
What the announcement does not settle
The sources are Google’s own release note and one news report, so the descriptions of what the agent can do are the company’s claims. Nothing published so far measures how reliably it finishes multi-step tasks, how quickly it works, or how often it asks for approval. The release notes give no usage limits per plan, no list of supported apps, and no examples beyond the general description. TechCrunch’s list of tasks is the only guide to what Google expects people to use it for.
For a Mac user on a paid Gemini plan in the US, the practical change is that repetitive file and form work can be handed to an agent that runs in a window of its own, with the user able to halt it at any moment. For everyone else, the feature remains a preview of what Google plans for the desktop app.
Sign Up for the Daily AI Pulse
One email a day. All the stories that matter.
By signing up, you agree to our Terms of Use and acknowledge the data practices in our Privacy Policy. You may unsubscribe at any time.
Alex CarterAI News & Big Tech CorrespondentAlex Carter writes for Aiholics, keeping readers updated on the fast-paced world of AI and Big Tech.



