Upload a voice sample → AI clones the voice → Generate speech or singing in that voice
Upload 10-30 photos → AI trains a custom LoRA → Use it for generation