Feature Requests

Could you add X and Y axis and Image rotation input fields next to the image size ones? I am physically disabled and use my voice to Control my computer. This makes using some Controls using a mouse difficult

Add some new tools for drawing,such as an ereaser (for cuting off an existing image) and fill bucket tool..so after merging your drawing you can restyle it in many creative ways.Thanks

When we use the Frame to add extra content to a photographic image, the lighting of the original image changes, so that the original image integrates well with the outline and edges, but the fact of adapting the lighting in the original image affects the details of the faces, sometimes causing blurring and obscuring. It would be great if there were a slider or option, to keep the original image intact, and to actually adapt the outpainting to the original photo and not the original photo to the outpainting and lighting of the photo, by means of a control in the frame's menus. Thank you in advance, for your consideration.

Would it be possible to integrate the prompt extraction feature which exists in the standard UI into the API ? (FR/ est il possible d'intégrer la fonction d'extraction de prompt dans l'API ?)

I’ve really enjoyed using Recraft, however, when it comes to making watercolor clipart, I’ve found the current prebuilt models a bit lacking. I love creating my own unique watercolor designs, but the models often add unwanted backgrounds or outlines—even when I request a white background or ask to exclude these features. I usually get my best results with the V3 Raw model, but even that doesn’t quite nail the look I’m after. I’m looking to create soft watercolor artwork with soft edges, without any harsh outlines. It would be amazing to have a dedicated watercolor clipart model that naturally produces images without backgrounds or outlines. This would give me and others the freedom to craft exactly the style we envision.

Flip the American Flag: I've repeatedly used an imported illustration of a boy wearing a backpack facing directly away from frame looking over a horizon of mountains. I say give me an "almost identical" vector illustration with just the text of "holding an American flag but the flag is flying upside down" - all i can get normally is just a "wisp" of a hint of red and white stripes in the sky or some type of flag that normally is just a corner popping out of the side of the boy and its hit or miss that its even an American flag. in the 1 in 20 chance that i get a semblance of an American flag it is ALWAYS right side up. i then follow with the "close" image and say "almost identical" and just enter in "flip the flag over upside down" or even more detailed "put the starfield at the bottom of the flag" and it will NEVER EVER EVER EVER flip the effing flag! Your AI SUCKS. If i can't get the simplest of prompts to work i'm dumping my sub. This is so VERY VERY VERY frustrating!!! Oh! BTW - the BOY is NEVER directly facing away from frame like the original image! WTF is up with that too?!? and with MULTIPLE reprompts instructing to have the subject/boy facing "DIRECTLY AWAY FROM FRAM" your AI will NEVER create a subsequent image in proper perspective!!! Loaded question: did you just finish training your AI three weeks ago? or was it four? (Oh gee... can you tell i'm more than peeved? Jesus, I'm PAYING for this!)

I also take this opportunity to make a second message concerning this time the mobile version which I have the impression has been completely abandoned despite the addition of the ICON recently added among the styles, personally I use a lot of graphic design like Punk graphic and pop graphic that I do not find in the mobile version do you think it is possible to do something about this? I would also like to talk about the magnificent option that you added during Valentine's Day "Valentinize", Since it was deleted a few days ago I add this term directly in my prompt but I have the impression that it is not necessarily the same result that I had when the option was added. Thank you in advance once again for the effort you give to your subscribers. Looking forward to reading you soon ♡ Kayline.

Image quality is too bad. I don't want to have zombie images. This is trash.

It is great to see the no text option within the projects when creating images, however despite adding no text and negative prompts for text/signature/watermark; I find 10 to 15% of the generated images have them. How can I have no text completely from API generated images? Thanks

I am having trouble getting it to follow building parts exactly-example: windows and shapes on the Micheal Graves engineering complex at University of Cincinnati. I don't want it to be creative on the structure itself--just control of the treatment of the existing structure.