I found some amazing AI art on civitai
retained text
Note that this wasn't created using just prompts. It was made in ComfyUI, an advanced node-based AI workflow suite associated with an ecosystem of thousands of open-weight local models such as StableDiffusion, Qwen, Z-Image, FLUX, assorted style LORAs, ControlNet adapters, OpenPose, and many others. It lets you specify the CFG, step count, VAE encoders and CLIP prompt conditioning encoders, etc. I included pictures of some workflow examples in the end. Civitai.com and HuggingFace have the biggest collections of them, and if you have an nVIDIA RTX graphics card or one of the other GPU devices now available for AI use, you can run the AI models locally on your own computer and learn to make stuff like this. There is quite a learning curve as you will be working directly with AI models, not the prompt-only toys most people are familiar with. This, however, represents the current state of the art tools and techniques used by advanced AI graphics designers. Tools like Photoshop are also used in conjunction, though I believe these were all fully AI generated. ChatGPT introduced me to this and has been teaching me to use ComfyUI and related tools. I have also created MCP tools to link OpenAI CODEX to ComfyUI so that it can take direct control of the ComfyUI interface and help me automate the process of workflow creation. If you are interested in doing advanced image generation with AI, this is the stuff to use. Credit goes to the original creators of these images on Civitai. I did not create any of these.
document fetched 2026-10-24 00:21 · canonical https://reddit.com/r/ChatGPT/comments/1xd58j1/i_found_some_amazing_ai_art_on_civitai