Mountain View, California, April 18, 2026, 08:32 PDT
Google’s Gemini Mac app has pushed the company’s flagship AI assistant onto Apple desktops, turning a web-first chatbot into a native macOS tool that can sit closer to daily work. Michael Friedman, group product manager for Gemini App, said the app is built as a native desktop experience “designed to live right where you work.” Blog
The timing matters because AI companies are racing to make assistants part of the computer itself, not just another browser tab. The Verge’s latest Installer column framed the shift plainly: new AI apps are now arriving for computers, including Gemini for Mac and OpenAI’s Codex update.
Google is also catching up. TechCrunch reported that OpenAI and Anthropic have had Mac apps for some time, while Google’s new app gives Gemini users a desktop shortcut and lets them share what is on screen, including local files, for more specific answers.
The app can be opened with Option + Space, with a full Gemini chat window available through Option + Shift + Space, MacRumors reported. It also supports window sharing, Dock and menu-bar access, image creation with Nano Banana and video generation with Veo.
Google Vice President Josh Woodward said in a post cited by Business Standard that the app was built by a small team in less than 100 days and was “100% native Swift,” referring to Apple’s programming language for Mac and iPhone apps. Until now, Mac users mainly reached Gemini through the web or Chrome. Business Standard
OpenAI moved in the same direction a day later with a major Codex update. The company said Codex can now operate a user’s computer by seeing, clicking and typing with its own cursor, while also adding an in-app browser, image generation, memory and more plugins.
Anthropic is already pushing Claude deeper into desktop workflows. Its download page describes Claude’s desktop app as bringing Chat, Claude Cowork and Claude Code into one place, while release notes say Claude Cowork became generally available on macOS and Windows through the Claude Desktop app on April 9.
That leaves Google with a narrower first version. Gemini for Mac appears built first for quick access and screen-aware help, while OpenAI and Anthropic are pressing harder into agents that can do work across apps. A desktop AI agent, in plain terms, is software that can act on files, browsers or apps after a user gives it permission.
But the path is not clean. Google’s Gemini for Mac page says users must share a window for context and enable Accessibility access if they want Gemini to read full browser pages, a system-level Mac permission that could give companies and privacy-minded users pause.
The wider bet is clear enough. Whoever owns the desktop prompt may get the first shot at emails, spreadsheets, charts, code and web pages before a user opens a rival app. Google has put Gemini into that race; now it has to prove Mac users want another assistant watching the workspace.
