
Your Mac just gained a second pair of hands. On November 19, 2026, Google switched on computer use inside the Gemini app for Mac, and the idea is simple to grasp. You describe a job in plain words. Gemini does the clicking, typing, menu diving and form filling for you.
The clever part is where the work happens. Gemini runs in an isolated background window with a live preview inside your chat. Your main screen and your cursor stay completely free. You can keep typing an email, designing a slide or taking a call while Gemini grinds through the boring part in the background.
This is not a chatbot that only answers questions. It is an agent that operates real apps and real files. It can open your folders, read a PDF, pull numbers from a live dashboard and build something new from what it finds. Google lists four starting examples: batch color grading, shipping quotes pulled from PDFs, decks built from live data and resale listings staged from photos in a folder.
There is a catch, and it is a narrow rollout for now. Computer use is rolling out in English in the United States, inside the Gemini app for Mac, for Google AI Plus, Pro and Ultra subscribers. You need an Apple silicon Mac running macOS Sequoia 15.0 or later, plus a personal Google account. Plans start at $4.99 a month, according to MacRumors, though one report notes that Google support pages currently name only Pro and Ultra, so check your own account before you pay for an upgrade.
Google is not alone in this race. Anthropic, Meta and OpenAI all shipped agents that work on their own computers in the autumn of 2026, and Apple is reportedly building a Mac native agent of its own. Gemini is Google’s answer, and it arrives with a twist the others lack: it works on the Mac you already own. Here are seven things it can do while you keep working.
1. It Sorts Your Messy Files While You Keep Working

Every Mac owner knows the folder. Downloads, screenshots and half finished projects piled into one place, with names like final-v2-real-final. Gemini can take that folder and turn it into order while you do something else entirely.
You tell it what good looks like. Sort these photos by date. Put every PDF invoice into a folder by month. Move the screenshots out of Downloads and into a folder called Screenshots. Gemini reads the files, opens Finder in its background window and moves things around, one click at a time, exactly as you would.
The win is not the sorting itself. It is the fact that your screen never changes. Older automation tools grabbed your cursor and locked you out until they finished. Gemini works in its own window, so you can answer emails or edit a document while the tidy up runs behind you. A live preview in the chat shows each step as it happens.
Start small. Hand it one folder, check the result, then hand it a bigger job. File moves are easy to undo when you catch a mistake early, and much harder to untangle after a thousand files have moved.
2. It Color-Grades a Whole Photo Shoot in One Go

Photographers lose hours to the same repetitive edits. The same exposure tweak, the same color fix, applied to two hundred photos one by one. Gemini can run that job across your favorite design software while you cull the next shoot.
Google names batch color grading as one of the first things to try. You describe the look you want, point Gemini at the folder, and it opens the app, applies the grade and saves each file. The work happens in the background window, so your main screen stays free for the creative decisions only you can make.
This matters because color work is fiddly and boring in equal measure. The sliders are small, the menus are deep, and the payoff for doing it by hand is tiny. An agent that can drive the app for you turns a lost afternoon into a ten minute check at the end.
Keep your expectations honest. Gemini follows your instructions, it does not invent your taste. Give it a reference photo or a clear description of the look, and check a few finished files before you let it loose on the full set.
3. It Reads Your PDFs and Fills Out Shipping Forms

Shipping quotes are a special kind of torture. The form wants a weight, three dimensions and an address, and every one of those facts lives in a different PDF. Gemini can pull them together and fill the form for you.
Here is how Google describes it. Gemini extracts weights, dimensions and addresses from local PDFs to fill out multi-box shipping quotes. It opens the files, finds the numbers, switches to the web form and types each value into the right box.
The same trick works anywhere a form repeats. Expense claims, supplier forms, registration pages. Any job where the facts sit in your files and the boxes sit on a website is a candidate. You describe the end result, and Gemini handles the copying and pasting.
Watch the first run closely. Numbers are the one place a small slip becomes a real problem, and a wrong weight can change a shipping price. The live preview lets you spot a bad value before the form is sent, and Gemini pauses for approval before key steps.
4. It Builds Your Presentation From Live Dashboards

A deck built from live data usually means a morning of screenshots. Open the dashboard, capture the chart, paste it into a slide, repeat for every region. Gemini can do the rounds for you and build the presentation in Google Slides.
Google’s example is specific. Pull metrics and screenshots from live dashboards across multiple regions and build a presentation. Gemini visits each dashboard, grabs the numbers and the visuals, then assembles the slides.
The payoff is freshness. A deck assembled by hand starts going stale the moment you paste the first chart. A deck Gemini builds in one pass carries the latest numbers from every source, because it collected them minutes before you present.
Give it a clear structure first. Tell it which regions to include, how many slides you want and what each slide should show. An agent with a vague brief builds a vague deck, and a precise brief is the difference between a draft you can present and a draft you have to rebuild.
5. It Stages Your Old Clothes for Sale While You Relax

The resale pile is the classic job that never gets done. The clothes are photographed, the prices are unresearched and the listings are unwritten. Gemini can work through the whole pile while you do something better with your evening.
Google’s example runs like this. Gemini analyzes clothing photos from a local folder, researches pricing, logs the item details and stages listings on resale platforms. It looks at each photo, works out what the item is, checks what similar pieces sell for and drafts the listing.
The trick is the combination. A chatbot can write a listing, and a photo app can show you the pictures, but neither can do both jobs and connect them. Gemini moves between your files, the web and the listing page in a single task, which is exactly what makes this kind of chore so slow for a human.
Keep the final say on price. Researched pricing is a starting point, not a verdict, and your item’s condition is something only you can judge. Let Gemini stage the listings, then review each one before it goes live.
6. It Turns a Rough Sketch Into a 3D Model

The most eye catching demo in Google’s launch post starts with a child’s butterfly sketch. Gemini takes the drawing, models it and rigs it in Blender, the free 3D app, turning a crayon idea into a moving model.
This is the creative workflow Google wants you to try first. Batch color grade photos in your design software or set up complex 3D scenes. The agent opens the app, navigates the menus and builds the scene step by step, while you watch the preview.
New desktop connectors make this possible. In the Gemini app, the Customize and Connectors area now lists desktop apps such as Blender and Autodesk Fusion alongside existing Git and Obsidian integrations, according to 9to5Google. You choose which apps Gemini may use, and it works inside them.
Creative apps are where an agent earns its keep, because the menus are deep and the clicks are endless. Setting up a 3D scene by hand can take an hour of small moves. Handing that hour to Gemini and spending your time on the idea instead is the whole point of the feature.
7. It Asks First and Lets You Watch Every Move

Handing an AI the keys to your computer sounds alarming, and Google knows it. The safety design is built around three controls: you choose which apps Gemini can access, you watch every step live, and you approve key actions before they happen.
The pause button matters most. Gemini can be stopped at any time with a single click, and it pauses on its own before sensitive steps such as logins and payments. It will not push through a checkout or a password screen while you are not looking.
Set the boundaries before your first task. Open the app permissions, allow only the apps the job needs, and keep the live preview open while it works. Treat the first few tasks as supervised practice, not autopilot, and check the result before you hand over anything with money attached.
There are limits worth knowing. The rollout covers personal Google accounts for users aged 18 and over, and the feature can make mistakes, as every agent can. Keep backups of anything important, start with low risk jobs, and let the guardrails do their job.
Computer use turns Gemini from a chat window into a coworker. It does the clicking, typing and form filling that eats your afternoons, and it does it in a window you never have to look at unless you want to. The seven jobs above are only the starting examples Google published. The real test will be the jobs you invent for it.
The timing is no accident. Anthropic’s Claude Duet, Meta’s Muse and OpenAI’s Dots all arrived in the autumn of 2026, each giving an AI its own computer to work on. Apple is reportedly building its own agent for the Mac, according to job listings reported this week. Google’s move brings the same idea to the Mac you already own, for the price of a subscription many people already pay.
Try it on a small job first. One messy folder, one batch of photos, one form. Watch the preview, check the result, and decide for yourself whether the hands are steady enough to trust. If they are, the most boring hour of your day may be the first thing you hand over.
Download the app at gemini.google/mac to get started. Computer use is rolling out now in English in the United States. Your Mac has a second pair of hands now. The only question is which boring job you hand over first.
