Google is rolling out an update to its Gemini app for macOS that adds voice-first assistance without requiring users to switch to a separate chat window. The new feature lets people trigger Gemini dictation system-wide by holding the Function (Fn) key, then turning spoken requests into finished text directly in the active app at the cursor. Google says the transcription removes common verbal fillers such as “ums” and “ahs,” and focuses on producing polished text where the user is working.
In addition to dictation, Google offers an opt-in “screen-aware reasoning” mode intended to give Gemini access to relevant on-screen context before handling more complex tasks. According to reporting from multiple outlets, this can include actions such as understanding what is shown on screen and, for example, turning highlighted notes into an executive summary and inserting rewritten text back into the document.
Overall, the update emphasizes quicker voice-to-text workflow across desktop windows, while providing optional screen context for tasks that may benefit from understanding what the user is viewing.
