🔞 ADULT: Just another ai app - HD Photos!
Just Another AI App
Packed with features to bring your creative ideas to life, for freeeeee:
-
Generate images and 3D models
-
Advanced features: inpainting, LoRAs, textural inversions, Controlnet and img2img.
-
3D model generation, so you can bring your flat 2D images to life!
-
Quality of life stuff:
-
Style presets selection.
-
Inpainting Mask editing.
-
Easy Lora and TI selection thru CivitAI.
-
3D model viewer
-
Background removal
-
How It Works:
-
AI Horde: Text-to-image gen service. No account setup needed; I’ve included a shared key loaded with kudos (thanks to the awesome volunteer network behind AI Horde, though wait times can be long).
-
HuggingFace: You’ll need your own free key, which gets you about 200 H100 GPU seconds daily. Select from any text-to-image models
-
3D Model Generation: Powered by Microsoft TRELLIS, hosted on my personal (potato) PC. It’s free but might be slow during busy times.
Why Should You Try It?
This app is perfect if you’ve been curious about AI-generated content but don’t have a gaming PC or the budget for pricey AI tools.
Give it a try, and let me know what you think! Feedback is always welcome. This is a project for the community, so I’m open to ideas for improvement or contributions if you’re a developer.
Let’s make creativity accessible to everyone.
Github Sauce code
Development log
- v 1.0 LaunchDec 15, 2024
Comments
Log in with itch.io to leave a comment.
Hey, a bit off-topic but I had an amazing AI expression morpher bookmarked from you and I see you took it offline. Any plan to bring it back? Or if not, is there anything similar you could recommend?
The PC hosting it died so I’m working on a new alternative. Right now you can definitely use nano banana (from Google) to do expression changes, it does it very well.
Hi, please check this I hope you can integrate and crate API
https://perchance.org/ai-text-to-image-generator
I dont think perchance has official API support
Lora search function works strangely.
It's looks like civitai changed something. I don't know exact rule, so I gonna show example:
Lets say that I want this LORA DetailedEyes_XL (https://civitai.com/models/120723/detailedeyesxl).
If I search by this name "DetailedEyes_XL" in civitai I got it right in first place. (and 239 other result, at this particular moment).
But If I use their API to get list of LORAs in json (which is also used by app), like this:
https://civitai.com/api/v1/models?limit=100&types=LORA&nsfw=true&query=DetailedE...
(since limit cannot be bigger than 100, you have to use "cursor" from metadata to goto next page)
It didn't occur until last page:
But it will work properly if you search for it with name without special characters ("detailedeyesxl"):
https://civitai.com/api/v1/models?limit=100&types=LORA&nsfw=true&query=detailede...
Of course in this way you cannot search for LORAs with space in name. It is strange.
My guess is that this API give response in ascending order. So the best results are very last and since app is getting only first 20 results of search it looks like it is getting something random unless you got search result with like one or few element.
Strangely adding "sort" option to api url doesn't help:
https://civitai.com/api/v1/models?limit=100&types=LORA&sort=Highest%20Rated&nsfw...
Lora search function doesn't work. İt always shows the same ones no matter what i type (also on textual inversions)
Nvm.
It works now
Thank's. It's cool.
Hey, Is there a way to get to raw images generated by app? I mean they are stored locally in browser memory, aren't they?
if you click on an image thumbnail theres a download button top right, theres also a download workspace button in the sidebar
Oh, I see now that I didn't explain myself clearly. I got some automatic browser update and it looked like the app lost all data.
I thought that, since I didn't delete anything, this information must be in some browser memory, but disconnected from a particular website.
It's no big deal, I lost things from that day before that I didn't download. From that point I always make backups with "Download Workspace".
What is happening with the generator? It has been sending me an error when generating the image for a few hours now.
should be fixed
It was just working again while I was creating a custom .json for my images, and the site is down again 🥲
The site is very good, thank you very much and we hope it is restored
Sorry about that, by default uses my aihorde key for image generations. Because I offer this for completely free, its used up too quickly. The key burns through 500k horde kudos a day >.<
If the generations are stopped it means my key ran out of kudos and I’ll need to recharge it.
Hey bro , this is one of the best tools i ever seen , where do i find it when you stop providing it here ? can i get the subscription somehow ? thanks for reply
happy new year
there are other free AIHorde apps:
https://tinybots.net/artbot
https://aqualxx.github.io/stable-ui/
However if you don’t have your own horde token then you’d be limited to anon generations which is slower and has limited max size when the horde is under heavy load. For my app you are using my token by default which is loaded with kudos.
You can go on aihorde discord to get free kudos https://discord.com/invite/3DxrhksKzn
How can I use advanced options like ControlNet Type - OpenPose?
I am new to AI and I tried to understand what it is from online tutorials, but the options I saw there are different from those in the application.
I tried to upload an image from OpenPoses in "ControlNet Image", check "Source Image is Control Map", but generation with this setting gives only errors.
to use controlnet, first select the controlnet option, then upload an image. In this example, the controlnet is openpose and the image shows the pose you want
Do not toggle on the Source Image is Control Map, that option is only for when you actually have the controlmap.
It seems like it just didn't work with the model I was testing - the Nova Anime XL.
I will investigate, but it doesn't look like it was a matter of the settings I had on that model.
ah yea I don’t think controlnet works for SDXL models(on ai horde)
Small bugs I've encountered:
In addition, when we change Advanced --> Express, the settings are not reduced to what we see in Express, the prompts remain even though they are not in the text field and e.g. the img2img settings remains.
Awesome.
I've been wanting to get into AI for ages, but I don't have a good enough graphics card. I stumbled upon this project by accident, it's awesome.
interesting!
Good job with this one!
Thanks, I’m glad you like it!