Google Launches Windows Search App And Mac Gemini Client For Desktop

Google released two new native desktop applications on April 15, 2026: a search app for Windows that debuted after seven months of beta testing, and a standalone Gemini AI interface for macOS built entirely in Swift. Both apps allow users to access Google's services directly from their desktops without opening a browser.

The Google app for Windows, available immediately, activates by pressing Alt + Space and displays a compact search interface that floats above other applications. Users can search the web or, with authorization, their local files and applications. The app includes Google's AI Overviews and AI Mode features, along with a Lens button that lets users highlight portions of their screen to initiate searches. The Windows version requires Windows 10 or 11 and currently operates only in English; users can access it without logging in, though logged-in accounts unlock additional features.

The macOS Gemini app, Google's first standalone desktop client for its AI assistant, launches with Option + Space and provides access to nearly all web-based Gemini features. The app can reference open windows for context when answering questions, includes file uploads, notebook creation, and access to Deep Research and Canvas tools. It also provides access to Google's image, video, and music generation models. The Mac app was built in less than 100 days and is coded entirely in Swift, according to CEO Sundar Pichai.

Google has chosen to distribute the Gemini app outside the macOS App Store, requiring users to download and install a DMG file directly from Google's website. The company made this decision despite the app being available in all regions and languages where Gemini is supported. There is currently no native Google search app available for macOS.

The Windows search app represents a significant expansion of Google's desktop presence. The company began testing the application in beta form last September, initially encountering technical limitations that prevented app updates; users had to uninstall and reinstall new builds during the beta period. The official release resolves these issues and adds the screen-context feature that allows users to share either a single window or their entire screen as context for searches.

Context

Google has historically relied on web browsers as the primary interface for its search and AI products, but desktop applications have become increasingly common across the technology industry. Microsoft offers native Windows apps for its Copilot AI assistant, while Apple provides native applications for most major services on macOS. The decision to build separate applications for Windows and macOS reflects divergent user bases and technical requirements—Windows users receive a search-focused tool, while Mac users get an AI-centric application.

The Windows app's seven-month development cycle from beta to release is typical for consumer software, though the beta period's technical constraints highlight the challenges of building cross-platform desktop applications. The macOS Gemini app's construction in less than 100 days using Swift, Apple's native programming language, suggests Google prioritized rapid deployment using established development frameworks rather than cross-platform tools.

What's Next

Google indicated that additional features for the Gemini Mac app are in development, though the company did not specify which capabilities or a timeline for release. The absence of a macOS search app suggests Google may be testing user demand for desktop search functionality before committing to a Mac version; if the Windows app gains significant adoption, a comparable macOS application could follow. The decision to distribute Gemini outside the App Store may face pressure from users accustomed to centralized app distribution, potentially influencing how Google handles future desktop application releases.