This gist converts Turkish audio to English audio.

Turkish_2_Eng.py

import Algorithmia, speech_recognition as sr, pyttsx

# Algorthmia Api key and end point
client = Algorithmia.client('simCM+BqJzKSw/iMrQsR/OTRwVE1')
algo = client.algo('translation/GoogleTranslate/0.1.1')

# pyttsx engine config
speech_engine = pyttsx.init('sapi5')  # see http://pyttsx.readthedocs.org/en/latest/engine.html#pyttsx.init
speech_engine.setProperty('rate', 150)


def speak(text):
    '''
        speaks the text(english) provided to it.
    :param text: text provided in english -- str
    :return: speaks using the speaker -- audio output
    '''
    speech_engine.say(text)
    speech_engine.runAndWait()


def listen():
    '''
        no arguments , records audio from the microphone and converts it to text.
    :return: text as interpretted from the audio(input voice recieved)
    '''
    # obtain audio from the microphone
    r = sr.Recognizer()
    with sr.Microphone() as source:
        print("Listening....say something...")
        audio = r.listen(source)
    print("... audio recorded ...")

    # 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)`

        # for English
        # print("Google Speech Recognition thinks you said in English: -  " + r.recognize_google(audio, language="en-US"))

        # for Turkish
        resp = r.recognize_google(audio, language="tr-TR")
        print("Google Speech Recognition thinks you said in Turkish: -  " + resp)
        return resp
    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))
    return None


def translate(text):
    '''
        translates text from english to turkish
    :param text: text to be translated to English
    :return: translated text
    '''
    input = {
        "action": "translate",
        "text": text
    }
    resp = algo.pipe(input).result["translation"]
    return resp


def main():
    audio = listen()
    resp = translate(text=audio)
    # print(resp)
    speak(resp)


if __name__ == "__main__":
    main()