Really impressive demo — the emotion and expressiveness feel far closer to a real human than most TTS I’ve tried. Curious: how do you handle consent and IP when cloning voices from short samples (10s)