Convert images to JSON requests, including dimensions and detailed information.
Want a different result? Change just a few parts
AI finds the swappable parts (subject/style/scene) — fill in what you want, keep the rest
