Granola launched a Mac voice-typing beta on October 29, extending its meeting-note software into everyday writing. Speech becomes edited text in the application already in use. The feature removes fillers, adds punctuation and supports hands-free dictation. It requires microphone and Accessibility permissions, and spoken instructions are entered as text rather than carried out.
Artificial Intelligence··Midday
Dictation reaches the active application
Granola, the meeting-note software company, launched voice typing for Mac applications on October 29.[1], [2]
Its beta turns speech into edited text in the field where the cursor is already placed, covering messages, documents and prompts. The feature is separate from meeting transcription: it does not start a new meeting or add dictation to one already running. It is included in every Granola plan.[1]
A shortcut controls speech capture
Users grant microphone and macOS Accessibility permissions before dictating. Accessibility access supports detecting the shortcut, obtaining application context and pasting the result. Users hold the shortcut to start dictation and release it to finish processing and paste the text. Double-tapping starts hands-free mode, which ends with another press or the on-screen control. Each session lasts up to 20 minutes. Cancellation during speech or processing discards the text without pasting.[1]
Speech becomes text without executing instructions
Granola removes fillers and accidental repetition, adds punctuation and arranges paragraphs and lists. A spoken question or instruction remains text rather than triggering an answer or an action. Language settings allow automatic switching while preserving each spoken segment in its original language. Users can choose a microphone. Computer audio is muted during dictation by default to reduce pickup from music or video. Windows and mobile support are forthcoming.[1]