Speech to text on a Mac, without uploading anything

macOS has built-in dictation, and for a sentence in a text field it is fine. Kana is for the rest of it: dictating into any app, transcribing recordings you already have, and getting text back that reads the way you would have typed it.

How it works

Hold a key anywhere in macOS and talk. When you let go, the text appears in whatever app had focus. Recognition runs on the Neural Engine on your Mac, so it works with the wifi off and there is no account to make.

Recordings, not just live speech

Drop an audio or video file onto the queue and you get plain text, SRT, VTT or Markdown back. There is no length limit and no per-minute charge, because transcribing on your own machine costs us nothing to run.

Where the built-in dictation stops

macOS dictation needs a field that supports it, has no file transcription, and cannot fix a name it keeps mishearing. Kana works in every app, takes files, and lets you register the words it gets wrong.