According to CoinMeta, Google has announced the launch of a global voice input feature for Gemini, for, and macOS. Users can start speaking by holding down the fn key in any application, and the transcription results will be directly inserted at the current cursor position. The system automatically removes colloquialisms such as "um" and "ah," recognizes changes in speech during transcription, and organizes the content into paragraphs or lists. With the optional screen inference feature enabled, Gemini can also read text, files, images, and documents selected by the user. Users can use voice commands to have the system rewrite content, adjust tone, extract summaries, or generate and modify images, with the results being written back into the current application. This new feature is being gradually rolled out globally, with initial support only for English.