Skip to content
Theme
OnDevice Tools

AIAI Tools & Apps

Gemini Mac computer use lets the app control desktop apps

Google is adding computer use to the Gemini app for Mac, allowing it to click and type across local desktop applications and websites. The feature runs in a separate window and includes controls for logins, payments and sensitive actions.

Mac desktop showing Gemini operating apps in a separate background window

Key takeaways

  • Gemini for Mac can perform clicks and typing across desktop apps and websites.
  • The assistant runs in a separate background window with a live preview in chat.
  • Google says Gemini will pause for logins and payments and ask before sensitive actions.
  • Users enable it by placing @Computer use in a prompt and choosing Desktop connectors.
In this article
  1. What changed
  2. How it works
  3. Example tasks
  4. How to enable it
  5. Why it matters

What changed

Gemini Mac computer use is coming to the Gemini app for Mac, according to a 9to5Google report published on 19 November 2026. Google says the update will let the assistant work across desktop apps and websites.

The feature expands the app's Task function, which was previously called Gemini Spark. Before this change, Task could reach local files only when used with connected services such as Gmail and Google Calendar.

Computer use is aimed at multi-step work. Instead of simply reading a file, Gemini can move through an app, select options and enter information as part of a wider workflow. In simple terms, it means the assistant can operate an app's visible controls, rather than only reading data through a connector.

How it works

Illustration: AI, Gemini logo, Google logo, computer and the labels Computer Use, Desktop Apps, Gemini

Google says Gemini performs the work in an isolated background window. A live preview appears inside the chat, so the user can see what the assistant is doing. The main display and pointer stay available for normal use.

This design is meant to let people keep working on their Mac while Gemini handles repetitive steps. The assistant can perform clicks and typing for the user, move through menus and complete forms. Because it works in a separate window, it does not need to take over the user's active pointer.

Safety controls are included. Users can halt Gemini immediately with a single click, and they choose which apps the assistant can use. Google says it will pause when a login or payment is required, and it will ask for confirmation before sensitive actions. The live preview gives users a way to follow the task before it reaches a sensitive step.

Example tasks

Google has outlined several tasks that could benefit from computer use. In creative workflows, the assistant could apply colour grades to batches of photos in design software or help create complex 3D scenes.

It could also reduce manual form entry. For example, Gemini could pull details such as weight, size and address from PDF files and use them to complete multi-box shipping quotes.

Another example involves live data. The assistant could gather data and images from live dashboards in multiple regions, then create a Google Slides presentation.

For resale, it could review clothing photos stored in a folder, check prices, record details and prepare listings on resale platforms. Google presents these as ways to manage repetitive parts while keeping the screen and cursor available. These examples show the feature is not limited to one app. It is intended for work that combines files, forms, dashboards and design tools.

How to enable it

To start a task, users place @Computer use in their prompt in Gemini for Mac. Google says the option appears in Customize, under Connectors.

The connectors area will show new Desktop options next to existing Git and Obsidian connections. Blender and Autodesk Fusion are named as examples of desktop software that can join the list.

The prompt marker signals that the task should use the new capability. This makes the feature opt-in for each request, rather than something that runs without the user asking for it.

Why it matters

Many daily tasks cross boundaries between apps and websites that do not connect. Google says computer use is meant to manage repetitive parts in the background, saving time while the user keeps control of the screen and cursor.

For Mac users, the feature could make Gemini more useful for practical work, not just questions and connected app data. It also shows Google expanding Gemini's role from answering requests to carrying out tasks across local software. The update points to assistants that can operate visible controls, with users able to watch and stop the process.

Sources

  1. Gemini app for Mac adds computer use to control desktop apps9to5Google

Tags

Spotted a mistake? See our corrections policy. Read how News by OnDevice works in our editorial policy.