app.py

import os
import torch
import gradio as gr
from TTS.tts.configs.xtts_config import XttsConfig  
from TTS.api import TTS

# ✅ Accept Coqui License Automatically
os.environ["COQUI_TOS_AGREED"] = "1"

# ✅ Allow `XttsConfig` in PyTorch's safe globals
torch.serialization.add_safe_globals([XttsConfig])

# ✅ Initialize XTTS Model (CPU-only)
tts = TTS(model_name="tts_models/multilingual/multi-dataset/xtts_v2", progress_bar=False).to("cpu")

print("✅ Model loaded successfully!")
print(f"✅ Supported Languages: {tts.languages}")  # ✅ FIXED

# ✅ Function for Voice Cloning
def generate_cloned_voice(text, reference_audio, language):
    output_path = "output.wav"

    if not text.strip():
        return "Error: Please enter some text.", None

    if not reference_audio:
        return "Error: Please upload a reference audio file.", None

    try:
        print("🔄 Processing voice cloning...")
        tts.tts_to_file(text=text, speaker_wav=reference_audio, file_path=output_path, language=language)
        print("✅ Voice cloning complete!")

        return output_path
    except Exception as e:
        print(f"❌ Error: {str(e)}")
        return f"Error: {str(e)}", None

# ✅ Create the Gradio Interface
interface = gr.Interface(
    fn=generate_cloned_voice,
    inputs=[
        gr.Textbox(label="Enter Translated Text"),
        gr.Audio(label="Upload Reference Audio", type="filepath"),
        gr.Dropdown(tts.languages, label="Select Target Language", value="en")  # ✅ FIXED
    ],
    outputs=gr.Audio(label="Generated Cloned Voice"),
    title="Free Voice Cloning API",
    description="Upload a sample voice and input text. Select a language, and the system will generate the text in the same voice.",
    allow_flagging="never",
    concurrency_limit=1  # Prevents multiple processes from running at once
)

# ✅ Launch the Gradio App
interface.launch()