yt2text

Python library for extracting text from a YouTube video in a single command, using OpenAi's Whisper speech recognition model. It doesn't use disk, performs everything in memory.

https://pypi.org/project/yt2text/

Installation:

pip install yt2text

Whisper requires ffmpeg to be installed in your computer. Check Whisper's requirements https://github.com/openai/whisper#setup

Usage:

You'll only interact with the get_text function. It takes a YouTube URL as an argument and returns the text as a string.

import yt2text

text = yt2text.get_text("https://www.youtube.com/watch?v=fLeJJPxua3E")
print(text)

Optional Arguments:

model: Set Whisper model (tiny,base,small,medium or large). Check here for details: https://github.com/openai/whisper#available-models-and-languages

Defaults to "base" which should be good enough for most cases. The first time you use a model, it will be downloaded first.

verbose Set True to print each step of the process. Defaults to False, it only prints if there is an error.

Usage with optional arguments

import yt2text

text = yt2text.get_text("https://www.youtube.com/watch?v=fLeJJPxua3E", model="medium", verbose=True)
print(text)

Contact

Raise an Issue in this Github repo (preferred, it sends a notification to my phone) Or mail me at [email protected]