Upload an image and Claude reverse-engineers a text-to-image prompt to recreate its look — with a style, lighting, camera and composition breakdown, plus Midjourney, Flux and Imagen variants. It recreates a similar image, never a pixel-perfect copy.
Yes — free and no sign-up. Upload an image and get the prompt plus variants.
No. It gives you a prompt that recreates a similar look and feel, not a pixel-perfect copy — image models are non-deterministic, so results will vary.
No. It's built to describe style and composition generically, not to clone a real, identifiable person's likeness or a trademarked logo — please don't use it for that.
Each model responds to different phrasing and parameters (for example Midjourney's --ar aspect flag). The tool tailors the prompt for each so you get better results in the one you use.