app.py

import streamlit as st
import outetts
from scipy.io.wavfile import write
import tempfile
import os
from pydub import AudioSegment

# Initialize model configuration
model_config = outetts.HFModelConfig_v1(
    model_path="OuteAI/OuteTTS-0.2-500M",
    language="en"  # Supported languages: en, zh, ja, ko
)

# Initialize the interface
interface = outetts.InterfaceHF(model_version="0.2", cfg=model_config)

# Streamlit UI
st.title("OuteTTS Speech Synthesis")
st.write("Enter text below to generate speech.")

# Sidebar for reference voice
st.sidebar.title("Voice Cloning")
reference_audio = st.sidebar.file_uploader("Upload a reference audio (any format)", type=["wav", "mp3", "ogg", "flac", "m4a"])

# Function to convert audio to WAV format
def convert_to_wav(audio_file):
    temp_audio = tempfile.NamedTemporaryFile(delete=False, suffix=".wav")
    audio = AudioSegment.from_file(audio_file)
    audio.export(temp_audio.name, format="wav")
    return temp_audio.name

if reference_audio:
    ref_audio_path = convert_to_wav(reference_audio)
else:
    ref_audio_path = None

text_input = st.text_area("Text to convert to speech:", "Hello, this is an AI-generated voice.")

if st.button("Generate Speech"):
    with st.spinner("Generating audio..."):
        # Generate speech with reference audio
        output = interface.generate(
            text=text_input,
            temperature=0.1,
            repetition_penalty=1.1,
            max_length=4096
        )
        
        # Save the synthesized speech to a file
        output_path = "output.wav"
        output.save(output_path)
        
        # Play the audio in the Streamlit app
        st.audio(output_path, format="audio/wav")
        st.success("Speech generated successfully!")

# Clean up temporary files
if ref_audio_path:
    os.remove(ref_audio_path)