Skip to content
Media MillDownload

How to transcribe video on a Mac without uploading it

Local transcription keeps the media and speech analysis on your Mac. You need a compatible Apple-silicon machine, a downloaded speech model and enough disk space for the source plus generated transcript and caption files.

Check the source before you transcribe

  • Use a clean copy when you have one.
  • Confirm that speech is audible above music and room noise.
  • Know the language being spoken.
  • Keep enough free storage for a copied source and its outputs.

Transcribe the file locally

  1. 1

    Install a local transcription engine

    Media Mill’s first setup downloads Parakeet v3, about 480 MB. Whisper is available as another local engine and downloads when first used.

  2. 2

    Add the local video

    Import the file. Media Mill copies it into a new Library item and leaves the original location untouched.

  3. 3

    Run Transcribe

    Use the Transcribe Action alone or the Transcript & overview Workflow when a short overview would be useful too.

  4. 4

    Check names and technical terms

    Open the timestamped Markdown transcript and compare uncertain passages with the recording.

  5. 5

    Keep the output you need

    Use Markdown for reading, timed JSON for structured work, or SRT and VTT for captions.

What stays local

Once the required models and media engines are installed, already-imported media can be transcribed without a network connection. Fetching a link and downloading those dependencies still require the network.