Gemini for macOS rolling out voice control and Gboard Rambler-level transcription - 9to5Google

What the report says

9to5Google reports that Google has begun rolling out advanced voice features for Gemini on macOS, following a preview at I/O 2026 in May. The update is described as letting users speak into any desktop window and have Gemini either enter refined dictated text or act on material visible on the screen. The rollout is global, currently for English-language users, with additional languages planned, according to the report.

The first feature is an upgraded dictation mode. 9to5Google says Gemini can turn spoken input into cleaner text by removing filler words and interpreting corrections made while speaking, then placing the resulting text wherever the cursor is active. The site compares the experience to Gboard Rambler, a transcription capability associated with upcoming Gemini Intelligence phones.

The second feature extends beyond transcription and requires users to opt in by enabling Gemini reasoning in the app’s settings. With screen context, Gemini can work from highlighted files, images, documents or text to summarize information, rewrite selected writing, generate visuals or revise existing images. Examples cited in the report include summarizing pet medical documents into an email, converting notes into an executive summary, or generating a dark-mode version of an illustration.

Users can activate the feature by long-pressing the Fn key anywhere in macOS or by using a new screen-sharing button in the Gemini prompt box, which brings up a floating waveform control. 9to5Google says users need version 1.88 of Gemini for macOS. The update matters because it moves Gemini further into system-level desktop workflows, where voice input, screen context and generative tools can be combined inside everyday Mac apps.