A visual-first AI image and video generator that combines subject, scene, and style references.