app.py

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

# ---------------- CONFIG ----------------
# On Hugging Face Spaces, add your Groq API key under Settings → Secrets → GROQ_API_KEY
client = Groq(api_key=os.getenv("GROQ_API_KEY"))

# Store last few messages for context
memory = []

# ---------------- LLM BOT ----------------
def llm_bot(prompt):
    # Use last few turns of context for continuity
    chat_history = "\n".join([f"{r[0]}: {r[1]}" for r in memory[-6:]])

    completion = client.chat.completions.create(
        model="llama-3.1-8b-instant",
        messages=[
            {"role": "system", "content": "You are a friendly and helpful AI voice assistant."},
            {"role": "user", "content": f"{chat_history}\nUser: {prompt}"}
        ],
    )
    return completion.choices[0].message.content.strip()

# ---------------- MAIN FUNCTION ----------------
def speech_to_speech(audio):
    if audio is None:
        return "Please speak something!", None

    # Step 1: Transcribe speech to text using Groq Whisper
    with open(audio, "rb") as f:
        transcription = client.audio.transcriptions.create(
            file=("speech.wav", f.read()),
            model="whisper-large-v3",
        )
    user_text = transcription.text.strip()

    # Step 2: Generate LLM response
    response = llm_bot(user_text)
    memory.append(("User", user_text))
    memory.append(("Bot", response))

    # Step 3: Convert LLM response to speech
    tts = gTTS(response)
    with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as temp_audio:
        tts.save(temp_audio.name)
        audio_out = temp_audio.name

    # Step 4: Return text + audio
    return f"🗣️ You said: {user_text}\n\n🤖 Bot: {response}", audio_out


# ---------------- GRADIO INTERFACE ----------------
with gr.Blocks(title="Groq Voice Chatbot 🎙️") as demo:
    gr.Markdown("## 🎧 Groq-Powered Voice Assistant")
    gr.Markdown("Speak naturally — it listens, thinks, and replies with speech using LLaMA-3.1 and Whisper on Groq!")

    mic = gr.Audio(sources=["microphone"], type="filepath", label="🎙️ Speak Here")
    output_text = gr.Textbox(label="🧠 Transcription + Bot Response")
    output_audio = gr.Audio(label="🔊 Bot Voice Reply")

    submit_btn = gr.Button("💬 Talk")

    submit_btn.click(
        fn=speech_to_speech,
        inputs=[mic],
        outputs=[output_text, output_audio]
    )

demo.launch()