app.py

import whisper
import gradio as gr


def get_text_from_mp3_whisper(mp3_file: str)->str:
    model = whisper.load_model("base")
    # options = whisper.DecodingOptions(language="en", without_timestamps=True)
    result = model.transcribe(mp3_file)
    return result.get("text", "No text found"), result.get("segments", {})
    
    
 
gr.Interface(
    title = 'OpenAI Whisper Transcribe audio files to text', 
    fn=get_text_from_mp3_whisper, 
    inputs=[
        gr.inputs.Audio(type="filepath")
    ],
    outputs=[
        "textbox", "json"
    ],
    live=True).launch()