Google's Gemini AI is expanding its capabilities on Mac with a new voice mode that allows users to dictate requests directly into any application. This feature bypasses the need for a separate chat window, integrating AI assistance seamlessly into the user's workflow.

Gemini's intelligent dictation mode transcribes speech at the cursor, automatically filtering out filler words like "ums" and "ahs." Additionally, a screen-aware reasoning option enables Gemini to understand on-screen content, allowing it to summarize highlighted notes or perform other complex tasks based on visual information. Users can activate these voice controls by holding down the Function (Fn) key.

This new voice functionality builds upon the native Gemini app released for Mac in April 2026. While Apple offers its own dictation and writing tools, Gemini's key differentiator is combining dictation, AI editing, and screen awareness into a single voice-activated shortcut.

The feature is rolling out globally in English for the Gemini app for macOS, with more languages planned for 2026. Google has not yet specified broader privacy or compatibility details.