Your Voice Memos library is a graveyard of good ideas: meeting debriefs, lecture recordings, 2am shower thoughts. Turning them into searchable text takes about five minutes to set up — and nothing has to leave your Mac.
Step 1: Get the files
If you record on iPhone, your memos already sync to the Mac via iCloud. Open the Voice Memos app on macOS and drag any recording straight to your Desktop (or directly onto your transcription app). Each one exports as a standard .m4a file. Recordings from other sources — dictaphones, WhatsApp voice notes, Zoom audio — work the same once you have the file.
Step 2: Drag into a local transcription app
In Sotto, drop the files into the Import window (or press Cmd+Shift+I). Supported formats: .mp3, .m4a, .wav, .webm. Transcription runs on your Mac's Neural Engine using whichever model you've selected — Whisper for broad language coverage, Parakeet for maximum English accuracy. No upload, no account, no per-minute charges.
Step 3: Fix the inevitable weird ones
Some memos are recorded in a windy parking lot. For those, two features earn their keep:
- Re-transcribe with a bigger model: if the quick pass with a small model mangled it, one click re-runs the same audio through Large V3 Turbo or Parakeet.
- Custom vocabulary: add the names and jargon you actually say, so "Kubernetes" stops coming out as "communities."
Step 4: Actually find things later
Every transcript lands in Sotto's history with its audio attached — full-text searchable, replayable, re-transcribable. The memo graveyard becomes a searchable archive: search "pricing idea" and jump straight to the recording where you said it.
Bonus: skip the memo step entirely
Half of voice memos are really just notes you couldn't type at the time. On the Mac, Sotto removes the middleman: hold a hotkey, speak, and the text lands directly in Notes, Obsidian, or wherever your cursor is. See how to dictate into any Mac app for the full workflow, or our guide to offline audio transcription on macOS for more on models and formats.