Skip to main content

​
Overview

On a Mac, the Local privacy mode dialog lists macOS Speech Recognition (built-in) · Fastest as the default transcription model. It uses the system’s own speech recognition, so you do not download hundreds of MB or several GB before the first run.
macOS Speech Recognition in Local privacy mode
On-device transcription progress and captions

​
Pain Point

I want Local privacy mode on my Mac without waiting on a large transcription model. The computer already has speech recognition; I want it ready after install.

​
Key Features

  • Mac-only option macOS Speech Recognition (built-in), selected by default, marked Fastest
  • No model download
  • Whisper Tiny / Base / Small / Medium / Large and others stay in the list (about 64 MB to about 3 GB)
  • Stays on this Mac: no upload, nothing written to the cloud library

​
Highlights

Local privacy mode can start immediately on a Mac instead of trading disk space for a first transcript.

​
Use Cases

  • First launch of the desktop app
  • A laptop that cannot spare extra gigabytes
  • Meeting audio that must not leave this Mac

​
Tips

  • For multilingual or very long audio, compare with Whisper options in the same list
  • Turn off Visual summary, then turn on Local privacy mode; pick language and model in the dialog

Local Privacy Mode

Desktop Drag-and-Drop Upload

Custom Transcription Engine

Upgrade to Pro: higher-accuracy online summaries → Pricing