Google Challenges Apple Intelligence With New Mac Gemini Upgrade
Google's native Gemini macOS app adds a floating voice dictation interface triggered by a long-press of the Fn key.
GOOGLEGoogle is bringing a new tool to macOS that lets users chat instantly with its Gemini Mac app in any desktop window following a single keypress. By doing so, the company is launching a direct competitor to Apple Intelligence and ChatGPT on the Mac desktop.
A Major Upgrade For The Gemini Mac App
By default, long-pressing the Fn key launches a new Intelligent dictation tool that automatically transcribes everything you say, stripping out any "ums" and "ahs," and places the result directly into your document.
However, you can also opt in to what the company calls "Screen-aware reasoning," that lets Gemini understand what's on your screen.
In an example video, Google demonstrates how the user can speak long, complex instructions, optionally referring to files on the local computer, and have Gemini synthesize a response directly into a document or email. In this example, the user points Gemini at a desktop folder including a list of dietary requirements, a selection of restaurant menus and a company expenses policy and asks Gemini to draft an email highlighting recommended options.
Gemini for macOS can read entire folders of files at once without the user needing to open them.
Multi-File Context And Zero-Copy Dictation
The single example voice prompt includes statements such as:
"Hey Gemini, I'm organizing our team dinner for next Thursday at 6 p.m. Can you check the files to make sure options fit our budget policy…"
and
"Oh, I'm sorry, it should be Friday at 6:30 p.m. Also search Google Maps for two great boba spots nearby…"
Without opening a single file or app, Gemini parsed five local documents, automatically disqualified an expensive restaurant and adjusted the timeline according to the user's correction.
With Screen-Aware Reasoning enabled, the user can highlight text in any app and ask Gemini to act, for example asking it to "turn these notes into an executive summary," and the assistant will place the summarized text into the document.
These tools mirror some of the functionality we've seen demoed in the company's forthcoming Googlebooks, which use an intelligent Magic Pointer tool to provide similar context-aware AI features deeply integrated into the desktop. With this latest update, some of these features will now be available to the Mac.
Google Gemini Mac App Vs. Private Cloud Compute
By binding Gemini to the Fn key, the new tool can be operated system-wide, making it feel as though it's part of macOS itself, rather than a separate app.
Furthermore, the unique advantage of Gemini's new assistant is its ability to read files directly from the macOS Finder and output text straight into open windows without the need for copy and paste. ChatGPT Desktop can only output into a floating window, leaving the user to paste the results into their documents manually and neither it or Apple Intelligence can currently perform this sort of folder-wide multi-file synthesis direct from the chat.
The key disadvantage, however, is that the Gemini app offers a strictly cloud-based service, accessed via a personal Google account. This means you don't get the same level of security intrinsic to the Mac's on-device models or Apple's Private Cloud Compute service that guarantees privacy even from Apple.
By contrast, Gemini app users must be comfortable sharing data with Google in accordance with its Gemini Apps Privacy Policy that permits data retention and human review by default.
Intelligent Dictation and Screen-aware Reasoning come as an update to the Gemini Mac app, rolling out worldwide from July 29 in English, with more languages promised later in 2026.
Withdrawn App Code Hides Google's Gemini Prompting Techniques
Google Photos Code Exposes Video Editor Changes For Android Users
This article was originally published on Forbes.com
