AI Art Generator for Android

I would give it a 5*, but the result of transforming a reference image to a specific style is irrelevant. Here’s a list of functionalities I hope you will add in the upcoming 2.1 or 3.0 update: - Fix the error popup that shows up after every 2/3 generated images - GPT-3 based comprehension of prompts - Ability to choose between Stable Diffusion, Waifu Diffusion, Anything and Midjourney engines - Seed guidance - Add the Influence style and artist to the prompts, like for example, by Studio Ghibli, trending on Artstation, by Picasso, by Leonardo Davinci - Extended positive and negative prompts text limit - Inpainting (img2img) and outpainting options - An experimental diffuser training model feature or standalone app (TestFlight)

Seed:102788

Most of my results didn’t even include the basic items i described. I tried multiple attempts with one common item and then different modifiers and details/location/Art styles and Every result looked mostly the same. I tried longer/more detailed descriptions but was still disappointed. If you want random Art it may suit you but it didn’t give me much.

This AI makes very little effort to change anything when you want it to make something closer to your prompt. Want it to take more steps so it’d be more meticulous of matching your prompt? No changes. Give it a scale to be as close as your prompt as possible? Just one piece is changed. Either this AI has very little creativity or it’s just plain lazy. Not even worth it.

About 2/3 of the time, I get an error message. However, when it does actually generate an image, it’s pretty accurate in terms of the prompt I entered. The developer should consider changing the animated progress image (displayed while an image is processing). The flashing colors may be seizure-inducing—certainly painful to look at.