A quickly hacked up script to transcribe Apple Voice Memos using OpenAI's Whisper. I don't plan to support or substantially improve this.
Usage:
python transcribe.py
The script performs the following steps:
- Tries to detect the current username
- Looks at the default path for Apple Voice Memos
- Loads up Whisper's medium English model
- Transcribes all the voice memos in the Voice Memo folder
- Outputs each transcription to a separate file, along with a master transcription file containing all transcriptions
Prerequisites:
- Whisper
- Python 3
- Apple Voice Memos synced to your Mac