My journey and exploration into AI art began on Monday September 12, 2022 when I had the flippant idea that I could make a pithy Facebook page full of inspirational quotes plastered onto generated images… and the joke in my head quickly became a little more serious, and then quite a bit more serious.

Below are the AI art generators that I’ve tried, some facts on them (as of this writing, 10/7/2022), and a few of my opinions. I tried the same prompt on all of these, with default settings. Where I was forced to choose a preset, I chose the most photographic/realistic.

The prompt I used:

Beautiful Fairy Princess In The Style Of Lady Clementina Hawarden, Hyperdetailed

NightCafe

NightCafe is currently my favorite AI art generator. There are three algorithms and the ability to upload a starter image, select from a library of free images or use one of your own previous creations. The customization options are nearly endless, while still having plenty of starter and more advanced presets to start out with. They have active communities on Facebook (including an active group) and Discord.

The biggest drawback for NightCafe is that the initial generated images are basically thumbs (384×576) and cost credits to upsize. Because there are SO many options, and each generation costs at least a credit (4 images using Stable Diffusion are 2 credits, or between 5 and 20 cents depending on what size of package you buy and plan you’re on). I signed up for Pro/Hobbyist for $9.99 a month, which gives me  fresh credits per day, and have bought credit packages numerous times since I seem to burn through them so fast (especially with video generation)! 

They do offer come credits when you share your images on social media, and reach certain milestones. There’s a pretty active community on the site, with lots of badges to earn, and it’s fun to browse by newest and comment and like other peoples’ works (and sometimes evolve them, if they’ve left those settings open).

NightCafe offers very customizeable AI video generation using their Artistic Algorithm, and I was able to fine-tune to get 10 seconds at 8 frames per second.

To generated the 4 images pictured, cost me 2 credits. 

NightCafe is currently my favorite AI art generator because of all the above.

You can find my NightCafe profile here, as ReachingMyDreams

Art made with NightCafe

NightCafe Creation Settings

Text Prompts

“beautiful fairy princess in the style of Lady Clementina Hawarden, hyperdetailed”

Weight: 1   

Initial Resolution

Thumb

Runtime

Short

Seed

64855819

Overall Prompt Weight

60%

Sampling method

K_LMS

Pixelz.ai

Pixelz.ai recently released some major updates which make them one of my new favorite AI art generators!

They have fairly advanced controls on three algorithms, with image input, the ability to upload a starter image. I didn’t see any beginner presets, although they do have some starter prompts within the app.

Pixelz.ai runs on a resource time model. I bought this during the middle of the night while half asleep and forget how much I paid (I guess I’ll find out when my credit card bill comes in), but mine tops off to 300 seconds every night.

They have a Discord channel, and you can also generate AI images through its prompts. For some reason it’s not connected to my Pizelz profile, or didn’t connect automatically. The Discord channel seems to work similarly to MidJourney…. the web app or the phone app is far more user friendly, I think!

To generate the six images pictured, cost me 60 seconds. They give 30 seconds when you make your work public.

Images without upsizing are 384×576 pixels, the same as NightCafe. Come to think of it, Pixelz.ai is similar to NightCafe in more ways than it is dissimilar, and they have an iPhone app (NightCafe is a browser bookmark/website icon). 

Again, the biggest technical way that Pixelz.ai is dissimilar to NightCafe is that there are no presets or starter settings. There is also no AI movie generation (that I saw). There is not the ability to follow other users, comment on their creations etc.

I think that this app/software shows a ton of promise, and I hope that it persists (despite rather bad reviews on the iPhone app store).

Find my public profile on Pixelz.ai as ReachingMyDreams.

 

Art made with Pixelz.ai

Algorithm: Stable Diffusion

Sampling steps: 50

Aspect: portrait

Sampler: ddim

Cfg Scale: 7.5

Seed: 47069094

 StarryAI

StarryAI was previously my most favorite, and then my second favorite AI image generator. It’s very easy to use as a newbie, and moderately customizable.

The AI generated portraits are almost always attractive and mainstream looking.

As with Pixelz.ai and NightCafe, there are three engines/algorithms, and you can upload your own starter image.

Unlike Pixelz.ai and NightCafe, you cannot adjust your prompt weight (how much the final image should match the prompt) or how similar the end image should be to the picture you uploaded.

You can earn credits by sharing on Reddit, Twitter and Instagram. The 4 images to the left cost 1 credit. Unfortunately with the Argo algorithm (similar to NightCafe’s Stable and Pixelz’s Stable Diffusion), you cannot generate only one image – you’re stuck with all 4 whether you like that or not!

There is the ability to generate animation/movies from the Orion/art setting like NightCafe, but it’s limited and seems to only show the animation of the AI being generated (the various iterative steps).

StarryAI has basic community with the ability to like and follow other creators. I signed up for a 3 months Pro for $29.99, which includes a daily top off of five credits. The portrait oriented images without upsizing are 512×640 pixels, larger than NightCafe or Pixelz, and StarryAI offers free upsizing to 2048×2560 dimensions.

StarryAI is really good for beginners or people wanting to play with prompts and generate ideas. There are lots of presets that are easy to work with and which create attractive images. However, if or once you want to branch out and get a little more artistically different and weird, this app/software is constraining.

Find my profile on Starry.ai as ReachingMyDreams

 

 

Art made with StarryAi

Prompt

beautiful fairy princess in the style of Lady Clementina Hawarden, hyperdetailed

MethodArgo
Runtime50

The other AI generators and apps

I really got on quite a roll with downloading and trying lots of AI apps! I put the name of the generator in the caption of the images below, again, using the prompt “Beautiful Fairy Princess In The Style Of Lady Clementina Hawarden, Hyperdetailed”, and with as realistic/plain settings as possible. 

 

 

WonderAI is pretty amazing (and I created my ReachingMyDreams and first AI art with this app)… It doesn’t have the ability to generate larger images (for print), generate from an image, or do very much else… but with unlimited creation and a fairly cheap yearly plan ($29.99) and (currently) $39.99 lifetime plan, it is well worth it! This is the app I use when I’m falling asleep, trying to find the right words or just wanting to play around. 

DawnAI is similar to DreamArt, but without any presets. It’s super basic, but the realistic high definition photograph quality of these images is incredible! It’s $1.99 a week with no other plans. Worth a try!

AI Art… Yep, this is about as generic as the app’s name!

The Dream Art app only allows for 300 characters. It’s pretty basic with lots of the same kinds of presets other apps offer. It’s fun to play around with and the current lifetime deal is only $49.99

Unidream possibly has potential, but only exports images as squares. With a 2000 character limit, you could get pretty detailed. At $28.99 a year it’s okay for the price. No ability to upsize, either. I don’t think this image is in the style of Lady Clementina Hawarden one bit.

Midjourney AI generated image

MidJourney generated AI image via Discord. I was disappointed.

Wombo Dream is pretty awesome, too. The lifetime deal of $139.99 for unlimited images seems okay, and it offers a high degree of customizations. As with StarryAI, the video generated seems to be a process video and not true AI generated video. 

Wombo does not offer the ability to hide prompts in images you publish, and no community features. I think it’s another app that’s great for those just dipping toes in the water with AI art creation or just wanting to generate ideas. 

Stay Updated

Sign up for uplifting and surreal art in your inbox! Might occasionally contain a coupon. 

You have Successfully Subscribed!

%d