A tool to use the Whisper v3 model from OpenAI to convert audio in a Youtube video to text and export it as a SRT subtitle file. You can then use the SRT file with tools like asbplayer for sentence mining. All transcription is done locally.

It also support transcribing audio file locally and exporting to a plain text file.

Note: the youtube downloader may not work. You can download the audio file through other means and provide it to the tool.

Clone this repository, create a virtual environment and then

pip install .

Note: if you want to use GPU for inference, follow the official PyTorch installation to get GPU working.

Usage: yt2srt [OPTIONS] SOURCE

SOURCE: name of audio file or youtube URL

Options:
  -t, --output_type TEXT      Output format. text or srt, default text
  -o, --output_filename TEXT  filename of the output file. Default to
                              output.txt
  --help                      Show this message and exit.