SpeechRecognition==3.10.0
# openai-whisper
soundfile
ffmpeg