Transcribe lyrics locally on your Mac

Lyric Video Studio includes a guided local audio-to-text workflow for creating a first lyric pass from a song.

Download a model

The first time you use transcription, choose and download a supported Whisper model.

The model is stored on your Mac so it can be reused for later projects.

Start transcription

Choose the project audio and begin recognition. Progress can be cancelled if you need to stop.

Recognition runs locally using the installed model.

Review the result

The review step shows detected segments before they are added to the project. Use the confidence cues and playback controls to check uncertain words.

Correct wording here before importing whenever possible.

Group and adjust

You can change how recognized text is grouped into lyric lines, apply an offset, and optionally snap the imported timing toward detected beats.

Import safely

Choose whether the result should replace the current lyrics, append to them, or be imported as an alternate lyric track.

This makes it possible to experiment without losing an existing timing map.

Manage downloaded models

Model storage can be managed from the transcription workflow so you can remove models you no longer need.