Guide
Transcribing a file you already have
Most transcription services want the file uploaded and charge by the minute. This one reads it on your own machine.
The short answer
Open an audio or video file in Voz and it is transcribed on your Mac by the Whisper model inside the app.
There is no per-minute cost and no upload, so the length of the recording only costs you time, not money or privacy.
How it works
- Open the Transcribe tab and choose a file, or drop one in.
- Voz converts it to the format Whisper needs using AVFoundation — there is no ffmpeg to install.
- The transcript appears and is kept in your local history, searchable with everything else.
What it will read
Whatever macOS decodes natively: mp3, m4a and AAC, wav, aiff, caf, flac, and the audio track of mov, mp4 and m4v. So a voice memo, a podcast, a lecture recording, or a video you were sent.
It will not read ogg, opus, mkv or webm. Those are rare for recordings people actually need transcribed, but if that is your file, convert it first or use something else.
Where it is weaker than a transcription app
Worth knowing before you buy
Voz gives you the text. It does not produce timestamps, subtitle files, or speaker labels by voice, and it bundles Whisper base rather than a larger model — so on a noisy recording or heavy accents, a dedicated tool with large-v3 will do better. If your job is turning recordings into subtitles or finished transcripts, MacWhisper is built for that and this is not.
This is a voice keyboard that also reads files, which is the right tool when you have a recording and want to know what is in it, and the wrong one when the transcript itself is the deliverable.
Why not upload it
Transcription services process your file on their infrastructure, which is fine for a podcast and a real question for an interview, a medical note, a legal recording or anything covered by an agreement. Voz has no server, so the file stays where it is. You can disconnect entirely and it still works.