How felipevbrito/transcribe_videos works
Program Transcribe is a command-line tool that turns a YouTube link into a saved text transcript. The user supplies a link; the program retrieves video information, downloads WAV audio, transcribes it with Whisper, and writes the text under output/. The README also describes Docker isolation, but the graph focuses on runtime product behavior. The workflow is visible in main.py; external-service edges reflect its yt-dlp and Whisper calls.
Main entry points: main.py CLI coordinator
Components 10
CLI workflow
- CLI coordinator
main.py - Video metadata
main.py
Audio processing
- Audio downloader
main.py - WAV audio
main.py - Whisper transcription
main.py
Transcript output
- Transcript writer
main.py - Text transcript
main.py
Other components
- User
- YouTube
- Whisper runtime
Connections 11
- User → CLI coordinator: provides link
main.py - CLI coordinator → Video metadata: requests info
main.py - Video metadata → YouTube: extracts info
main.py - CLI coordinator → Audio downloader: downloads audio
main.py - Audio downloader → YouTube: downloads audio
main.py - Audio downloader → WAV audio: writes WAV
main.py - CLI coordinator → Whisper transcription: transcribes audio
main.py - Whisper transcription → WAV audio: reads audio
main.py - Whisper transcription → Whisper runtime: loads model
main.py - CLI coordinator → Transcript writer: saves text
main.py - Transcript writer → Text transcript: writes transcript
main.py