A more robust prompted oriented Image-to-image modifier/enhancer

The ability to upload a small to medium-sized image or photo and either edit or enhance it, or simply have fun with it, is highly desirable. For instance, one could upload a photo of their car parked on a street and use a prompt like "recreate this image but make the car fully engulfed in flames" or "jumping over the Grand Canyon." Detailed prompt-based edits and enhancements are just around the corner, I'm sure. Think of Topaz Gigapixel's redefine model that works according to the user's prompted direction but that actually works. The innovation your team has created here is truly groundbreaking and is helping shape the future of AI in general. Bravo.