Welcome to the ARTSMARTAI roadmap!

It would be interessting, to keep the charakter/s in different poses and backgrounds. So that only the scene chance. Like in a picture book or comic. In other words, creating a series of pictures with "actors" who are clearly recogniseable in each single picture.

The width/height sliders only give a few options of image sizes.  For webpage headers, I want closer to a 4:1 ratio, but the best allowed by the sliders is 2:1.  Allow sliders down to 128 or 256, and change the steps to that size (or increase the max to 2048 or 4096)

Suggestions to create web pages

It would be great to directly let the AI integrate a watermark (i.e., artist's signature) to an image. This would make the artworks really unique.

They are add-ons, like Outpainting.

Use Scribble - Also try to turn something you scribble to an image.

OpenPose detector - Annotates human keypoints such as positions of the head, shoulders, hands, etc. In some apps there is an editor to change the position of keypoints.

https://github.com/lllyasviel/ControlNet

An option to add negative prompts would be useful. 

The "Compare Models" feature lets you simultaneously generate and compare images from up to four top AI image generator models. This powerful tool streamlines your creative process by allowing side-by-side comparisons, helping you quickly identify the model that best suits your artistic needs.

SUGGESTIONS:

Mass selection of images for download.

Arrows for easy navigation through all generated images.

Drag and drop to easily organise creations.

I function in which you can bulk changes your images would be great. For example bluk delete or bulk change privacy setting.

UPDATE!!
Our new image organization system has been released, allowing you to batch delete, batch download, and bulk update image visibility. Learn more : https://docs.artsmart.ai/collection#69b3a06575914ffba257fc2f08f6115b.

I'd like to have a way to generate an image and then insert the text of a blog heading onto it.

When my team does this, we end up using presets with a certain font, color, size, etc. To keep the images looking similar. Everything comes out nicely branded. (See an example here: https://vrnews.io/virtual-reality-games/)

Would like to see:

• Ability to overlay text to an image after generating.
• Ability to save brand presets with font family, colors (rgba), background, relative sizes, etc.
• Ability to apply a color/gradient overlay to the whole image OR just behind the text with some padding. 
• Bonus points: If we can do all this via the API.

Thanks!

It would be interesting to be able to give names to characters, objects and scenarios, and then reuse them in different prompts.

Ex: @DarthVader and @Yoda playing 3D chess on @TheLittlePrince'sPlanet.

Add Rating Button for user can teach AI to know which is good or bad images.

Maybe Like or Dislike Button or 1 to 5 Star rating.

Most of websites are based on WordPress CMS.

It would be useful to provide a wordpress plugin in order to be able to generate images and store them directly in the media library of a wordpress WebSite, in order to illustrate articles or pages. 

Some areas of the website (e.g. generate button) cannot be easily accessed when using a browser on a mobile device. 

Maybe it's my computer...but I can barely see anything on the screen. The tiny scrollbars can barely be seen... the current color scheme and the tiny scrollbars are a nightmare when combined. Please make the scroll bars always visible (no disappearing scrollbars), please create a light mode option and please increase the width of the scrollbars so people can quickly click on them instead of struggling to find them in the dark. 

Please understand that users want to be efficient and take as little time as possible to create on your website... hunting for basic user features (like scrollbars) is insanely inefficient and creates a miserable user experience.

A clickable box for location where a person or object if applicable is preferred to be rendered like the camera preference command presets

Would love to see these graphic styles available on ArtSmart.ai. Please click the link below and if you agree, please vote on this change request :-)

https://huggingface.co/nitrosocke/Nitro-Diffusion

Bulk generate images based on prompt.

Contextually, I enjoy generating images, but dislike waiting for it to generate, then clicking generate again, and so on so forth. You could always add a warning about it taking a long time.

This may just fall under the "outpainting" that is in the works, but I'll throw it out here anyway. It would be extremely handy to be able to add a bleed to an image. We'll get images and artwork that we need to print edge to edge and we'll need to add a 1/4 inch bleed on all 4 sides. We don't want any new information or elements added to the image, just the image seamlessly expanded out enough that we can use the full image in print.

AI art is still a work in progress.  Sometimes it's hard to get exactly what you're hoping for.  

Being able to hire out small adjustments like add logo and text to an image would be helpful!

Various objects including brains and other organs that look correct.  To be able to make variations on these objects, and tweak images that are almost the way we want them, but just needs some small changes.

Midjourney have a blend images where you can mix two images with AI.
https://docs.midjourney.com/docs/blend
Many people search this on google because it's not available anywhere else and Midjourney doesn't provide an API.

So there is no other easy way to do that.

Here is an interesting concept: https://huggingface.co/spaces/lambdalabs/image-mixer-demo

The Ai doesn't add letters properly when you request so l, in my marketing artworks are important part of the art so could be cool to have this done by the AI.

Write Exactly what we request in an artistic way to go together with the background or other images.

Love the latest changes! However, it would be highly recommended to add Pagination to the image library and add a "Select All" option so you can bulk download images. For those of us with hundreds or thousands of images, it's impossible to manually select each image for download. And it's impossible to find a specific image through endless scrolling. Adding this feature would not only make things much easier for users, but it would allow us to purge our image library which would free up storage space on your network.

I am looking for a feature that allows you to upload an image and use certain parts (i.e., a character) of the image as a source to reuse in another image. Here is an example: I want to illustrate a book and for this I create a cover page with the main character of the book. Now I want to create another image of the main character within the book, but in which he/she is e.g. doing a different activity or is in a different situation, state of mind, etc.. However, the character should, of course, look exactly like the one on the cover. Prompt example for the cover: A girl, standing in a magical forrest surrounded by animals, golden hour, sun rays, fairy dust, watercolor painting In the next step I want to upload the cover image and be able to select the girl and let her do something else: Prompt Example: Use the girl from the source image. Let her sit at a wooden table besides her grandmother, watercolor painting, close up protrait and so on... I hope the example is understandable. If not, just ask me and I'll try to explain it better. If this function already exists, how can I use it?