YouTube Audio Transcription Tool

This tool automates the process of transcribing YouTube content by downloading videos' audio tracks and converting them to text using OpenAI's Whisper model running locally. The resulting transcriptions are saved as text files, which can then be further analyzed using ChatGPT or other tools.

Update 2025-04: You can use this tool for the purpose of just downloading the audio as well.

Requirements

• Python 3.8 or higher
• FFmpeg (required for audio processing)

Before you start

1. Create and activate a virtual environment:

Mac/Linux

python -m venv whisper-env
source whisper-env/bin/activate

Windows

python -m venv whisper-env
whisper-env\Scripts\activate

2. Install the required packages:

pip install -r requirements.txt

Transcription

Downlaod and transcribe:

python youtube_transcribe.py "https://www.youtube.com/watch?v=VIDEO_ID"

Advanced options:

python youtube_transcribe.py "https://www.youtube.com/watch?v=VIDEO_ID" [OPTIONS]

Options:
  --browser {chrome,firefox,opera,edge,safari,chromium}
              Uses cookies from specified browser, for paid content
  --delete-after                    Delete audio files after transcription

Download only

python download.py "https://www.youtube.com/watch?v=VIDEO_ID"

Download with low quality

python download.py "https://www.youtube.com/watch?v=VIDEO_ID" --audio-quality 32 --sampling-rate 16000

Transcribe files of audio/videos

python transcribe_from_files.py --path path_to_local_video_or_audio --model large --audio-quality 32 --sampling-rate 16000

Transcribe a folder of audio/videos

python transcribe_folder.py --path path_to_folder --model large --audio-quality 32 --sampling-rate 16000

The tool supports:

• Youtube single video URLs
• Youtube playlist URLs (except Watch Later playlists)
• Videos requiring authentication (via cookies file or browser cookies)
• Videos with non-ASCII titles (including Chinese characters)

Downloaded audio files are saved in the audio/ directory, and transcriptions are saved in the transcripts/ directory.