Voice Memos: recording, and why transcription needs a second setup step

Recording only needs ffmpeg. Turning a memo into a note needs Whisper as well, plus its own Settings switch, which can be off even when the Transcribe now button is visible.

Last checked against the app on · View as Markdown

Steps

  1. Open Voice Memos and press record. It captures as .m4a and stops itself automatically after thirty minutes if you forget it.
  2. Recordings are stored in their own folder next to your vault, not inside it, so a raw audio file never shows up in search or on the Map.
  3. To get a written note from a memo, press Transcribe now. This needs Whisper installed with a model configured, separately from ffmpeg, and the memo transcription switch in Settings turned on.
  4. A successful transcript is written as a new note with a Summary and a full Transcript section, linking back to the original audio file rather than copying it in.

How to tell it worked

A transcribed memo appears as a note under wiki/memos, and the original recording is still the file it points back to.

If it does not work

Transcribe now fails immediately even though the button is there.
The button being visible does not mean transcription is switched on. Check the memo transcription toggle in Settings; if it is off, transcription refuses outright and says so.
One memo fails to transcribe and you assume the whole batch stopped.
It did not. Each memo is attempted on its own, and one honest failure, Whisper missing or a write error, does not block the rest.
A batch was interrupted partway through.
At most one memo is left stuck mid-transcription, and it clears itself the next time you run the batch.