app.py

import os
import gradio as gr
import whisper
from gtts import gTTS
from groq import Groq

GROQ_API_KEY = os.getenv("GORQ_API_KEY")
client = Groq(api_key=GROQ_API_KEY)

# Load Whisper model
model = whisper.load_model("base")

def chatbot(audio=None):
    if audio is None:
        return "No input detected. Please provide an audio input.", None

    # Transcribe the audio input using Whisper
    transcription = model.transcribe(audio)
    user_input = transcription.get("text", "")

    # Generate a response using Llama 8B via Groq API
    chat_completion = client.chat.completions.create(
        messages=[{"role": "user", "content": user_input}],
        model="llama3-8b-8192",
    )
    response_text = chat_completion.choices[0].message.content

    # Convert the response text to speech using gTTS
    tts = gTTS(text=response_text, lang='en')
    tts.save("response.mp3")

    return response_text, "response.mp3"

def clear_inputs():
    return None, None, None

# Create a custom interface
def build_interface():
    with gr.Blocks(css="""
        .block-title {
            text-align: center; 
            color: white;
            background-color: #4CAF50; 
            padding: 10px;
            border-radius: 8px;
        }
        .gradio-row {
            background-color: #f9f9f9;
            border-radius: 8px;
            padding: 20px;
            margin: 10px;
            box-shadow: 0px 4px 12px rgba(0, 0, 0, 0.1);
        }
        .gradio-column {
            padding: 10px;
        }
        .gradio-button {
            background-color: #ff6347 !important;
            color: white !important;
            border-radius: 8px !important;
            padding: 10px 20px !important;
            font-size: 16px !important;
            border: none !important;
            cursor: pointer !important;
            box-shadow: 0px 4px 12px rgba(0, 0, 0, 0.2) !important;
            transition: background-color 0.3s ease !important;
        }
        .gradio-button:hover {
            background-color: #e5533d !important;
        }
    """) as demo:
        gr.Markdown(
            """
            <h1 class="block-title">Voice-to-Voice AI Chatbot</h1>
            """
        )
        with gr.Row(elem_classes="gradio-row"):
            with gr.Column(elem_classes="gradio-column", scale=1):
                audio_input = gr.Audio(type="filepath", label="Record Your Voice")
            with gr.Column(elem_classes="gradio-column", scale=2):
                chatbot_output_text = gr.Textbox(label="Chatbot Response")
                chatbot_output_audio = gr.Audio(label="Audio Response")

        clear_button = gr.Button("Clear", elem_classes="gradio-button")

        clear_button.click(
            fn=clear_inputs,
            outputs=[audio_input, chatbot_output_text, chatbot_output_audio]
        )

        audio_input.change(
            fn=chatbot,
            inputs=[audio_input],
            outputs=[chatbot_output_text, chatbot_output_audio]
        )

    return demo

# Launch the interface
if __name__ == "__main__":
    interface = build_interface()
    interface.launch()