shubhammehta5315/VOICEcloner ? reverse-engineered prompt

Build me a Google Colab notebook that lets me run a simple web app for multilingual text to speech and voice cloning.

I want to paste in text, choose or upload a voice sample, and generate speech that sounds like the reference voice. It should also support voice design and a stronger cloning mode if that’s available. Please make the experience easy for a non expert, with a clean Gradio interface and clear steps to run it in Colab.

Since this will be used in Colab, make it work smoothly on a T4 GPU if possible, and include a setup that keeps downloaded files and model cache in my Google Drive so I do not have to redownload everything every time. If you need to check current model or library setup details online, feel free to look them up.

Please make the notebook ready to open and run with minimal manual setup, and keep the instructions simple inside the notebook.