New top local AI image generator is here!

MOMOSS-ImageOpenMOSSAvailable in EU

Open-source 5B-parameter text-to-image model with strong knowledge, text rendering, infographics, and layout control via bounding boxes. Runs on low VRAM; creator installs it in ComfyUI, tests text-to-image and layout workflows, and notes it is text-to-image only (inpainting for editing).

K2Krea 2KreaAvailable in EU

Leading open text-to-image model, more than twice the size of MOSS Image. Creator says it performs somewhat better in initial tests but needs more VRAM.

ZIZ-ImageAvailable in EU

6B-parameter image model cited as MOSS Image's closest competitor; creator says MOSS has better understanding and flexibility.

COComfyUIComfy OrgAvailable in EU

Popular free platform for running open-source image, video and audio generators locally. Used to run MOSS Image workflows, including LoRA and inpainting setups.

HKHiggsfield KatanaHiggsfieldUnverified

Sponsor segment: AI video editing tool working inside Claude that recreates viral videos shot by shot with your own characters, products or footage.

Community LoRA for MOSS Image that applies an ink-splatter art style, shown in both text-to-image and layout workflows. Requires trigger words.

Community fine-tuned MOSS Image checkpoint that looks better and more consistent than the base model; drops into the diffusion models folder.

Turbo LoRA that cuts generation to 8 steps at CFG 1, making MOSS Image roughly three to four times faster, at the cost of negative prompt effect.

ATAI ToolkitOstrisAvailable in EU

Open-source toolkit for training LoRAs, now with MOSS Image support, and it also works on macOS.