Gemini for macOS now lets you hold the Fn/Globe key and speak into almost any Mac app. It can clean up dictation at your cursor by default; if you opt into reasoning, it can also rewrite selected text, summarize files or images, and generate or edit images using on-screen context.
Quick setup: use Gemini voice in any Mac app
- Download the official Gemini app for macOS and sign in.
- Open any app with a text field, such as Mail, Notes, Docs or a browser.
- Press and hold the Fn/Globe key, speak naturally, then release it.
- Gemini removes filler words, follows mid-sentence corrections and inserts formatted text at your cursor.
- For hands-free, longer dictation, double-tap Fn to start and tap it again to finish.
Compatibility: Google says the Gemini desktop app runs only on Apple Silicon Macs and requires macOS Sequoia 15.0 or later. The new voice feature is rolling out globally in English to Gemini for macOS users; availability can vary during rollout.
Turn on reasoning for rewrites, summaries and images
Basic intelligent dictation is enabled by default. More advanced tasks require an explicit opt-in:
- Open Gemini Settings.
- Select Speak to Window.
- Turn on reasoning.
- Highlight the text, document, PDF or image you want Gemini to use.
- Hold Fn and say what you want it to do.
With reasoning enabled, Gemini can use the visible screen context for more than transcription. Google’s examples include turning notes into an executive summary, extracting details from local files into an email, and generating a dark-mode version of an illustration.
Gemini for Mac shortcut cheat sheet
| Action | Shortcut |
|---|---|
| Speak to the active window | Hold Fn/Globe; release to submit |
| Long hands-free dictation | Double-tap Fn to start; tap again to stop |
| Open mini chat | Option + Space |
| Open full chat | Option + Shift + Space |
| Share the front window as context | Press both Command keys |
Google says these shortcuts can be customized in the Gemini app settings.
Useful voice prompts to try
- Clean dictation: “Write this as a concise client update with three bullet points.”
- Rewrite selected text: “Make this warmer and clearer without changing the facts.”
- Summarize a document: “Summarize the selected PDF into decisions, deadlines and action items.”
- Create an executive summary: “Turn these notes into an executive summary with a TL;DR at the top.”
- Edit an image: “Create a dark-mode version of this illustration while preserving its layout.”
Privacy and permission checklist
- Keep reasoning off if you only need cleaned-up dictation.
- Review the active window before invoking screen context; Gemini uses what is visible in the foreground window.
- Only grant Accessibility access if you need Gemini to read complete browser pages or files in a connected folder. Google directs users to macOS System Settings → Privacy & Security → Accessibility for this permission.
- Do not dictate passwords, financial account details, private client data or confidential health information.
- Verify generated summaries and rewrites before sending or publishing them.
FAQ
Is Gemini for macOS free?
Google describes the desktop app as available at no cost where the Gemini app is supported, although selected features require a Google AI subscription and feature availability varies.
Does Gemini for Mac work on Intel Macs?
No. Google’s current compatibility page says the app runs exclusively on Apple Silicon Macs and requires macOS Sequoia 15.0 or newer.
Why don’t I see Speak to Window yet?
Google announced a global English rollout on July 29, 2026. Staged rollouts can take time, so update the app and check Settings → Speak to Window. Google says more languages are coming later but did not provide dates.
Does voice dictation require screen reasoning?
No. Intelligent dictation is on by default. Screen-aware reasoning is an optional setting for tasks such as rewriting selected content, summarizing files and generating images.