How to transcribe audio or video on Mac

September 11, 2026 · 6 min read

A recorded interview, lecture, call, or voice memo is most useful when the audio and the words remain connected. The NotchLive transcription app for Mac can import an existing audio or video file, turn its speech into text locally, and create a saved session you can play, review, translate, and export.

The original file remains untouched. NotchLive does not upload it to a NotchLive cloud transcription service.

Before you import

• Install NotchLive 1.12 or later and finish speech-model setup.
• Choose the model and source language that fit the recording. Auto works across many languages; an explicit language can help when the file is mostly one language.
• Make sure you have permission to transcribe the recording. Consent and recording rules depend on the people, place, and purpose involved.

File import is a Pro feature. NotchLive requires macOS 14 or later; saved-transcript translation requires macOS 15 or later and a supported language pair.

1. Start from Home

Open the NotchLive window and choose Import from Home. Select the audio or video file you want to work with. Import is separate from live captioning, so you do not need to play the file aloud or route its sound through another app.

If live captions are active, NotchLive coordinates the work rather than making two speech-recognition jobs compete. Follow the state shown in the app before resuming another intensive speech task.

2. Let NotchLive create the saved session

Speech recognition runs on your Mac using the selected model. For a video import, NotchLive retains the audio needed for session playback instead of keeping a second copy of the video. The source file itself is not edited.

Transcription time depends on the recording length, model, Mac, language, and audio quality. Crosstalk, distant microphones, music, and heavy noise can reduce accuracy. Check names, numbers, and consequential wording against the recording.

3. Read and listen in the same session

Open the completed item from Sessions. Use the audio controls to replay the recording beside its transcript. Search, copy, share, or export from the session instead of managing a loose media file and text document separately.

If the retained audio is moved or deleted outside NotchLive later, the session explains what is missing and keeps the transcript available.

4. Review speakers on supported Macs

On supported Apple silicon Macs, imported recordings can receive local speaker detection. Labels begin as anonymous, meeting-local evidence—not real-name identification. Review the participant list and transcript before relying on speaker-specific analytics or sharing the record.

NotchLive does not create persistent voiceprints or recognize the same person across unrelated sessions.

5. Add language tabs without replacing the Original

Choose translation from the saved session and select a supported target language. NotchLive keeps the Original intact and creates a separate tab for each translation. Every language tab has its own search, copy, share, and export actions.

You can leave the session while translation continues. Higher-priority live speech work may pause the job; NotchLive resumes it when resources are available. Saved audio is not required—a text-only session can be translated too.

Choose the right Mac transcription workflow

Import when the recording already exists. Use meeting transcription on Mac when you want captions during the event, a reviewed participant record, and a saved session afterward. For speech playing live through another app, start with on-device speech-to-text on Mac.

In NotchLive 1.12, a live recording can optionally retain Compact audio at about 22 MB per hour or Lossless speech-quality mono audio. Audio saving is off by default.

Both paths lead to the same useful idea: one private session where the transcript, optional audio, speaker review, translations, and exports remain connected.