import speech_recognition as sr import gradio as gr def listen(audio): r = sr.Recognizer() sample_rate, audio_data = audio sample_width = audio_data.dtype.itemsize # with wave.open(audio, 'rb') as audio_file: # sample_width = audio_file.getsampwidth() audio = sr.AudioData(audio_data, sample_rate=sample_rate,sample_width=sample_width) try: text = r.recognize_google(audio, language='ur') return text except: return "Sorry could not recognize what you said" gr.Interface(fn=listen,inputs="audio",outputs="text").launch()