app.py

import os
import streamlit as st
from groq import Groq
from pathlib import Path
from pydub import AudioSegment

# ------------------------------
# Fetch API key from Secrets
api_key = os.environ.get("GrokAPI")
if not api_key:
    st.error("❌ Please set GrokAPI secret")
    st.stop()

client = Groq(api_key=api_key)

# ------------------------------
st.title("🎤 Audio → AI Text → Speech")

audio_file = st.file_uploader("Upload audio", type=["wav", "m4a"])

def truncate_text(text, max_chars=1000):
    """Limit text size for TTS"""
    if len(text) > max_chars:
        return text[:max_chars] + "\n\n⚠️ Answer truncated for TTS."
    return text

if audio_file:
    try:
        # ------------------------------
        # Convert to WAV
        audio_path = Path("input.wav")
        audio_segment = AudioSegment.from_file(audio_file)
        audio_segment.export(audio_path, format="wav")

        # ------------------------------
        # Transcribe audio
        transcription = client.audio.transcriptions.create(
            file=("input.wav", audio_path.read_bytes()),
            model="whisper-large-v3",
            response_format="text"
        )
        st.text_area("📝 Question", transcription, height=150)

        # ------------------------------
        # Generate AI answer (shorter)
        completion = client.chat.completions.create(
            model="llama-3.1-8b-instant",
            messages=[{"role": "user", "content": transcription}],
            temperature=0.7,
            max_completion_tokens=150,  # short answer
        )
        answer_text = completion.choices[0].message.content
        st.text_area("💬 AI Answer", answer_text, height=200)

        # ------------------------------
        # Truncate answer to safe length for TTS
        answer_text_limited = truncate_text(answer_text, max_chars=1000)

        # ------------------------------
        # Convert text → speech
        speech_path = Path("answer.wav")
        response = client.audio.speech.create(
            model="playai-tts",
            voice="Aaliyah-PlayAI",
            response_format="wav",
            input=answer_text_limited
        )
        response.stream_to_file(speech_path)
        st.audio(str(speech_path), format="audio/wav")

    except Exception as e:
        st.error(f"❌ Error: {str(e)}")