Skip to content

Free delayed real time transcription for youtube livestreams using wit.ai, ffmpeg and github actions

License

Notifications You must be signed in to change notification settings

dli-invest/fdrtt

Repository files navigation

Real time cheap as heck transcription service

Transcribe video codecov

Components

Github actions

open sourcing this content again for free computing power

Libraries

  • ffmpeg
  • wit.ai or speech recognization
  • youtube-dl
  • asyncio
  • one thread for wit.ai call

Python cmd tool to parse a youtube livestream and save it in 4:50 minute intervals to send to wit.ai for speech recognization (free to use).

Next time powell gives a speech I will be ready.

Use the wit-ai api.

https://medium.com/wit-ai/pure-audio-transcription-7372aa0bda7e

wit.ai blows with wav files, never rely on them.

pseduoCode

initialize youtube dl to get video

grab first element in video, various variables to track performance

video download complete

wit.ai logic in async run with discord posting and outputting txt file

grab next element in video.

should be all todo all this in vscodespaces


### Test Coverage


### Other

Yahoo finance livestream starts at 1 pm UTC, monday to friday, maybe skips some holidays.

About

Free delayed real time transcription for youtube livestreams using wit.ai, ffmpeg and github actions

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages