app.py

import gradio as gr
from stt import speech_to_text
from llm import ask_llm
from tts import text_to_speech

def voice_to_voice(audio):
    if audio is None:
        return "Please record audio first", "", None

    user_text = speech_to_text(audio)

    if user_text.strip() == "":
        return "Could not recognize speech", "", None

    ai_text = ask_llm(user_text)
    ai_audio = text_to_speech(ai_text)

    return user_text, ai_text, ai_audio

demo = gr.Interface(
    fn=voice_to_voice,
    inputs=gr.Audio(type="filepath", label="🎤 Speak"),
    outputs=[
        gr.Textbox(label="🧾 You said"),
        gr.Textbox(label="🤖 AI replied"),
        gr.Audio(label="🔊 AI Voice"),
    ],
    title="🎤 Voice-to-Voice AI (Free & Open Source)",
)

demo.launch()