import speech_recognition as sr

def startpy():

    # obtain audio from the microphone
    r = sr.Recognizer()
    d= ''
    while (d!='exit' and d!='quit'):
        with sr.Microphone() as source:
            print("Say something!")
            audio = r.listen(source)

    # recognize speech using Google Speech Recognition
        try:
            # for testing purposes, we're just using the default API key
            # to use another API key, use `r.recognize_google(audio, key="GOOGLE_SPEECH_RECOGNITION_API_KEY")`
            # instead of `r.recognize_google(audio)`
            data = r.recognize_google(audio)
            print("Google Speech Recognition thinks you said " + data)
        except sr.UnknownValueError:
            print("Google Speech Recognition could not understand audio")
        except sr.RequestError as e:
            print("Could not request results from Google Speech Recognition service; {0}".format(e))
        d = data

#startpy()