I want to setup and test a basic & simple "Chatgpt-4o Voice-To-Voice App". I already have setup ChatGPT-4o-preview model in Azure AI Foundry, and I already have the API Key & the endpoint. I just want "THE CODE FOR THE APP"

Question

I want to setup and test a basic & simple "Chatgpt-4o Voice-To-Voice App".

I already have setup ChatGPT-4o-preview model in Azure AI Foundry, and I already have the API Key & the endpoint. I just want "THE CODE FOR THE APP" (Microsoft has not provided the code to use, for setting up an app for this)

Answer 1

Hi there Santosh Kumar

Thanks for using QandA platform

you will need to integrate both speech recognition (converting voice to text) and speech synthesis in addition to using the ChatGPT-4 model through the Azure API.

cofde snipp for voice to voice app

import os
import openai
import azure.cognitiveservices.speech as speechsdk
import pyaudio
import wave
# Setup Azure Cognitive Services Speech API
speech_key = "YOUR_AZURE_SPEECH_KEY"
region = "YOUR_AZURE_REGION"
# Setup OpenAI (ChatGPT-4) API
openai.api_key = 'YOUR_OPENAI_API_KEY'
chatgpt_endpoint = "YOUR_CHATGPT_ENDPOINT"
# Speech-to-Text: Initialize speech recognizer
def speech_to_text():
    speech_config = speechsdk.SpeechConfig(subscription=speech_key, region=region)
    audio_config = speechsdk.audio.AudioConfig(use_default_microphone=True)
    recognizer = speechsdk.SpeechRecognizer(speech_config=speech_config, audio_config=audio_config)
    print("Say something...")
    result = recognizer.recognize_once()
    if result.reason == speechsdk.ResultReason.RecognizedSpeech:
        print("Recognized: {}".format(result.text))
        return result.text
    else:
        print("Speech could not be recognized.")
        return None
# ChatGPT-4: Send text to ChatGPT-4 and get response
def chatgpt_response(prompt):
    response = openai.Completion.create(
        model="gpt-4",
        prompt=prompt,
        max_tokens=150
    )
    return response.choices[0].text.strip()
# Text-to-Speech: Convert text response to speech
def text_to_speech(text):
    speech_config = speechsdk.SpeechConfig(subscription=speech_key, region=region)
    audio_config = speechsdk.audio.AudioConfig(filename="output_audio.wav")
    synthesizer = speechsdk.SpeechSynthesizer(speech_config=speech_config, audio_config=audio_config)
    synthesizer.speak_text_async(text)
    print("Speaking the response...")
# Main function to run Voice-to-Voice
def voice_to_voice():
    while True:
        # Step 1: Listen and convert speech to text
        user_input = speech_to_text()
        if user_input is None:
            break
        
        # Step 2: Send input to ChatGPT-4 for processing
        chatgpt_text = chatgpt_response(user_input)
        print(f"ChatGPT-4 Response: {chatgpt_text}")
        # Step 3: Convert ChatGPT-4 response back to speech
        text_to_speech(chatgpt_text)
# Run the app
if __name__ == "__main__":
    voice_to_voice()

Kindly accept the answer if this helps thanks.