Google Adds Speech Transcription to macOS Gemini App for Desktop Workflows

Google is rolling out a new tool in the macOS Gemini app that allows English-speaking users to transcribe speech and clean up filler words directly into any desktop app. Available now with more languages planned later, the feature also includes a screen-aware reasoning mode for summarizing files and rewriting text.

The updated desktop application arrives as part of a push by Google to integrate AI deeply into everyday computer workflows. Rather than keeping artificial intelligence confined to a browser tab or a standalone chat window, the tool operates right inside the windows where people are already typing, creating, and editing.

How Voice Transcription Works Across Desktop Applications

The new functionality is designed to function seamlessly without requiring users to copy and paste text manually.

To use the feature, a user hovers their cursor over a target window, holds down the Fn key, and speaks. The software then captures the audio and inserts the transcribed text straight into that active window.

Beyond basic translation of sound to letters, the system cleans up conversational filler. It automatically removes hesitation words such as “ums” and “ahs” while catching mid-sentence corrections on the fly. Google refers to this default configuration as intelligent dictation.

Screen-Aware Reasoning and Advanced Desktop Assistance

Transcribing speech is only part of the update. The macOS tool also incorporates a screen-aware reasoning mode that grants the assistant permission to examine on-screen content.

With screen access enabled, users can ask Gemini to summarize open files, rewrite selected blocks of text, or handle other document-based tasks directly on their display.

Apple Ecosystem Timing and Availability

The release strategy for this desktop update requires specific user interaction. At the same time, users retain the ability to opt out of the screen-awareness settings whenever they choose.

How to Use Gemini Live on Desktop (Google AI Studio Tutorial)

The timing of the launch places Google’s tool on macOS desks while Apple’s own refreshed Siri remains in beta testing and is not expected to roll out until September.

English-speaking users can begin using the speech transcription and screen-aware features immediately. Google has confirmed that support for additional languages will roll out at a later date.