Save and Share:
After months of development, our vision of a tool for generating pictures from text was rolled out into open beta.
The service is based on the Stable Diffusion neural network. Examples of images created by the neural network:
I won’t bore you with lyrics, although I’m very proud of the team (which, by the way, has grown, but more on that later), and I have something to tell, so here are the main features:
- You can generate as much as you want for free
- Our solution, unlike others, does not require “prompt engineering”; that is, you just need to write “cat”, select the category of the model, and you will get the best neural cat. The pictures above are obtained on the first try, by simple requests
- With Credits: you can generate privately, in priority, and 40 pieces each
- Without Credits: no priority queue, and can take ~15 minutes, we are slower than others in this regard
- NSFW is prohibited, this is monitored by a separate neural network
- You can send public results to your friends, the link will open without registration
- If you like some art, just click “re-create” and the text query and category selection will immediately be substituted
- The generator only works with English
- What categories are there now:
- tattoos (sketches),
- textures (for gamedev and 3D people),
- And more categories will be added.
- The page shows what other people generate
- Images are generated with a CC0 license, you can print, sell, do NFT, commerce, etc.
- Directly on the site, you can improve the result by upscaling (increase, reduce blur, etc.), for this there is an Enhance button
So, Let’s crash our servers: https://neural.love/ai-art-generator
(Registration is required, we do not send marketing emails if you do not check the box)
PS. This project would not have been possible without the research work of CompVis and Stability AI, these organizations are AI gifts to humanity