I’d love to see a feature that allows users to upload their own images for the AI to edit or enhance. Ideally, the AI would be able to detect the content of the image and apply prompt-based transformations accordingly.

Additionally, I’ve noticed that prompt interpretations are often inconsistent or vague, even when they are clearly worded. It would be great to see improvements in how the model understands and executes detailed prompts.

One more important thing: the AI frequently makes anatomical errors (like strange hands or missing faces), or simply avoids generating full faces altogether. This really breaks the realism and limits creative uses. Improving facial rendering and anatomical accuracy would take the platform to a whole new level.

Thanks for the great work so far – this tool has massive potential!