Google has added system-wide voice input to Gemini for macOS. Now, in any app, you can hold down the Fn key and start speaking. Gemini will transcribe your words directly at the cursor—and it’s smart enough to strip out “ums” and “uhs,” handle mid-sentence corrections, and automatically format your speech into clean paragraphs or lists.
If you opt in to screen reasoning, Gemini can also read what you’ve selected—text, files, images, documents. Then you can just talk to rewrite the content, adjust its tone, pull out a summary, or even generate and edit images. The results get dropped right back into the app you’re using.
The feature is rolling out globally to all Gemini Mac users, starting with English only.