April 18, 2026, 08:32 PDT—Mountain View, California.
Google has brought its Gemini AI assistant straight to Apple desktops with a new Mac app, transforming it from a web-only chatbot into a native macOS feature that sits alongside users’ work. “Designed to live right where you work,” said Michael Friedman, group product manager for Gemini App, describing the app as a fully native desktop experience. Blog
Timing is key as AI firms scramble to embed assistants directly into computers, instead of leaving them stuck in a web browser. The Verge’s current Installer column put it bluntly: fresh AI apps are landing on desktops, like Gemini for Mac and the latest Codex upgrade from OpenAI.
Google is moving to close the gap. According to TechCrunch, OpenAI and Anthropic rolled out Mac apps earlier, but Google’s just-launched app hands Gemini users a desktop shortcut and the ability to share on-screen content—including local files—for more tailored responses.
MacRumors says you’ll launch the app using Option + Space, or hit Option + Shift + Space to pull up the full Gemini chat. The software’s got window sharing, sits in both the Dock and the menu bar, and lets you make images with Nano Banana or generate video using Veo.
Google’s Josh Woodward, a vice president, told Business Standard that a small team managed to put the app together in under 100 days, building it entirely in Swift—Apple’s own programming language for Mac and iPhone. Previously, Gemini on Mac was only accessible via web browsers or Chrome.
A day after, OpenAI rolled out a significant Codex update. Now, according to the company, Codex handles a user’s computer—watching the screen, moving the cursor, clicking, and typing on its own. They’ve added an in-app browser, image generation, memory, and extra plugins as well.
Anthropic is moving fast to integrate Claude into desktop routines. The company’s download page pitches the Claude desktop app as a hub for Chat, Claude Cowork, and Claude Code, all bundled together. Release notes show that Claude Cowork rolled out broadly for both macOS and Windows via the Claude Desktop app on April 9.
So Google’s rolling out a more limited initial release. Gemini for Mac looks focused on speed and context—fast access, plus help that’s aware of what’s on your screen. OpenAI and Anthropic, though, are pushing further into agent territory, aiming for tools that tackle tasks across different apps. A desktop AI agent, put simply, handles actions in files, browsers, or apps once users grant access.
The process isn’t seamless. Google notes on its Gemini for Mac page that anyone wanting Gemini to access full browser pages will have to share a window for context and turn on Accessibility access—a Mac system permission that could raise privacy concerns for both companies and individual users.
The logic behind the push is simple: control the desktop prompt and you’re first in line for emails, spreadsheets, charts, code, and web pages—before the user even touches a competing app. Google’s Gemini has jumped into that contest. Next up: convincing Mac users they actually want another assistant peering over their digital shoulder.
