
Phonon-2, built in
A 164 MB speech model with 5.21 % word error on the Open ASR Leaderboard, running at 174x realtime. Read about Phonon-2
Detta turns speech into finished, formatted text in any app, with recognition running entirely on your Mac.


A 164 MB speech model with 5.21 % word error on the Open ASR Leaderboard, running at 174x realtime. Read about Phonon-2

Gluon removes fillers, repeated words and false starts before the text reaches your document.

A third tap saves a note with its recording and a screenshot, readable by AI agents you connect on the same Mac.

Audio and video files, from WAV and MP3 to MP4 and MOV, are transcribed by the same model.
macOS 15 or later · Sign in with Google required · 709 MB download, about 1.1 GB installed · beta