app.py

import gradio as gr
import pytube
import whisper



def get_audio(video_url):
    # Download YouTube video
    yt = pytube.YouTube(video_url)
    return yt.streams.first(only_audio=True).download(filename="tmpvideo.mp4")


def transcribe(video_url):
    # Transcribe audio
    model = whisper.load_model("base")
    return model.transcribe(get_audio(video_url))

iface = gr.Interface(fn=transcribe, inputs="text", outputs="text")
iface.launch()