AI Art Generation Handbook/List of local AI Art WebUI

The WebUI listed here are updated since June 2024.

There are some criteria that are needed so that this local AI Art WebUI are included inside:

1. They should be free (in terms of beer) and not paid (Therefore for example: Commercially available Adobe Firefly is totally out)
2. They do not need to include online registration (ala link to Google's account) or needed always on-line to be functions (like makeayo)
3. They should be notable enough for other user to take notice and promotions through words-of-mouth (or in this age, social media)

See also: List of AI Art Websites if you want to use it online or don't have the required hardware.

Logo	Author Name	Website Link	Remark
	Automatic 1111	A1111 Web UI	One of the most earliest and most versatile WebUI with external extensions built in. It is also one of the largest community support (based on Github stars)
	comfyanonymous	ComfyUI	Instead of typing text as programming input, it uses node programming to "program" as flow-state in text to image generation
	cmdr2	Easy Diffusion	Easy to install WebUI with one-click install
	lllyasviel	Forge	Optimisation from the A1111 fork which can produce faster image generations and also optimise the memory management to reduce OOM error message
	lllyasviel	Fooocus	WebUI that tries to emulate the ease of use ala MidJourney. Note: This is the recommended local WebUI for the beginners if they want to try AI Art Generations
	invoke-ai	InvokeAI Stable Diffusion Toolkit	It have have unified UI for inpainting and outpainting workflow.
	n00mkrad	NMKD Stable Diffusion GUI	This fork support for AMD GPU
	lllyasviel	omost	This fork using Large Language Model to describe descriptively about the images so that it can render the images in different part of images and combined together in single image
	Vladmandic	SD.Next	A fork from Auto1111 which can support multiple types of models , improved prompt parser, optimised for multiple platform and Lora enhancer for training.
	VoltaML	VoltaML	Recompiled ML library into high performance inference runtimes like TensorRT, TorchScript, ONNX and TVM meaning that it can generate double the speed from A1111

References

https://www.reddit.com/r/StableDiffusion/s/s8rqnqOE3N