AI News
AI News AgentProductGoogle2 min read

Gemini automates repetitive tasks on your Mac

Gemini adds computer use to its Mac app and can complete multistep tasks across files, applications, and web pages. The feature works in the background, shows every action, and lets you choose access permissions, pause processes, and approve important operations.

Gemini can now handle repetitive tasks on your Mac: it clicks, types, opens menus, and fills out forms for you while working with your local files, desktop apps, and web pages.

The computer use feature is coming to the Gemini app for Mac starting today. Instead of only answering questions, the assistant can run multistep processes from a written instruction in plain language.

For example, you can ask it to review a folder of photos, batch-edit them in your design software, and prepare the results. It can also extract weights, dimensions, and addresses from PDF files to fill out a shipping quote with multiple fields.

What it can do for you

Google has outlined several initial uses for the feature:

  • Apply color adjustments to many photos or prepare complex scenes in design and 3D modeling software.
  • Read data from PDF documents and transfer it to shipping forms.
  • Gather metrics and screenshots from live-data dashboards across different regions to create a presentation in Google Slides.
  • Analyze photos of clothing, research prices, record the details, and prepare listings on resale platforms.

The goal is to reduce processes that normally require dozens of clicks. Instead of switching between windows, copying information, and filling out each field manually, you describe the result you need and Gemini carries out the steps in between.

You stay in control

Gemini does not take control of the screen you are using. The feature works in an isolated window in the background, with a live preview inside the chat. Your main desktop and cursor remain available so you can keep writing, designing, or taking a call.

You can also decide which apps it can access. The preview lets you watch every action, pause the process at any time, and approve important operations before they happen.

That matters because these tasks can involve personal files, open accounts, and forms containing sensitive information. The automation does not work like an unsupervised command: you can review what Gemini is doing and stop it if something does not match your request.

The feature is rolling out globally in English in the Gemini app for Mac. To try it, you need to download the app from the official Gemini for Mac page.

The more important change is that AI assistants are starting to move from answering to acting inside the tools you already use. What you will need to watch now is how accurately they interpret instructions and how they behave when a task combines local data, applications, and web services.