TypeWhisper
DEGitHub

File Transcription

Transcribe audio and video files manually, or automate incoming files with Watch Folder.

Manual File Transcription

Use File Transcription when you want to transcribe existing audio or video files without recording a new dictation. Drag files into the drop area or choose them from the file picker.

Manual batches process only the files you add yourself. You can choose the transcription engine, model, task, and language behavior before starting the batch.

File Transcription settings with folder selection, transcription settings, output format, and source file deletion toggle

Watch Folder Automation

Watch Folder monitors a folder you choose in TypeWhisper settings. When a supported audio or video file appears there, TypeWhisper queues it and transcribes it automatically.

Watch Folder keeps listening for new files until you stop watching. This is useful for recorder exports, downloads, or other folders where new media files arrive repeatedly.

Output Files

TypeWhisper writes the transcript artifact first. Choose Markdown, plain text, SRT, or WebVTT as the output format.

For Watch Folder, transcripts are written to the selected output folder. If no output folder is configured, transcripts are written next to the source file in the watched folder.

Delete Source Files After Transcription

When this Watch Folder option is enabled, TypeWhisper deletes the original source audio or video file after the job finishes successfully and the transcript output has been written.

The source file is removed directly by the app. It is not deliberately moved to the macOS Trash.

If transcription fails, TypeWhisper keeps the source file in place so you can retry or inspect the problem.

Only enable this option when you no longer need the original files or you already keep another backup.

Basic Setup

  1. Open Settings > File Transcription.
  2. For manual batches, drop files into the file area or choose them from the file picker.
  3. For Watch Folder, choose the watched folder and, optionally, a separate output folder.
  4. Choose the engine, language behavior, output format, and whether source files should be deleted after successful Watch Folder transcription.