# A quick MOSS-Image news round-up - 14th November 2026 — retained record

> Retained reddit record from reddit:comfyui.

Canonical: https://highsignal.app/data/records/1dccb5535b1c8dfb

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:comfyui)  published 2026-11-14 16:03 ingested 2026-11-15 02:36

[original source ↗](https://www.reddit.com/r/comfyui/comments/1xw5y0y/a_quick_mossimage_news_roundup_14th_november_2026/)

## retained text



```
Another quick MOSS-Image news and goodies round-up, for those who may have missed some items. 

-> MOSS AnyPaint, offering... "arbitrary-mask inpainting, outpainting and image editing for MOSS-Image-Base."

https://huggingface.co/yijunwang2/moss-anypaint

-> A Layered Papercut style LoRA for MOSS-Image. The trigger is *ral-pliq* and a suitable prompt. A pleasing realistic craft style, but perhaps adding a simplification LoRA would give a more convincingly hand-made look? I also wonder if prompting for *backlit, diorama*  might not give a realistic shadow-box effect?  

https://civarchive.com/models/3049898?modelVersionId=3468199

-> A semi-real fine-art painting style for MOSS-Image SFT, which can still do toony faces - yet is also evidently able to ditch the anime eyes and do 'classic 1970s western fantasy-art' faces and eyes in a Ted Nasmith -like style. The backdrops also look excellent, so this is probably also worth trying for big fantasy landscapes? No trigger, but the maker suggests *traditional oil painting style, semi-realism* be used in the prompt.

https://civarchive.com/models/3051322?modelVersionId=3470027

-> Tiny Thing for MOSS. Emulating macro photography of fine painted miniature models and doll-art.

https://civarchive.com/models/2313869?modelVersionId=3470483

-> MOSS-Image Concept Sliders for ComfyUI. The high-angle slider initially looked useful, but on testing it has a tendency to elongate the head. This nodes-set also offers NAG, and an implementation of "Midjourney's *--sref* codes" for MOSS, which lets you apply styles by typing shortcodes which then call their full style-prompt from a .JSON styles file. Has workflows.

https://github.com/quinteroac/ComfyUI-MOSS-Concept-Sliders

-> ComfyUI-MG-Cache. A potential MOSS-Image speed-up, via enabling "multi-granularity caching" for MOSS.

https://github.com/KidiXDev/ComfyUI-MG-Cache

-> And finally, if you still need GGUF models then Abiray has a handy new set of 4-step MOSS-Image-SFT GGUFs at various quants.

https://huggingface.co/Abiray/MOSS-Image-SFT-4-steps-GGUF/tree/main

~ OLD POSTS ~

https://old.reddit.com/r/comfyui/comments/1xuze4s/a_quick_mossimage_news_roundup_13th_november_2026/ (see this post for link to old posts) 

https://old.reddit.com/r/comfyui/comments/1xixxjn/a_quick_mossimage_news_roundup_30th_october_2026/ (see this post for link to older posts) 

https://old.reddit.com/r/comfyui/comments/1xf2nwp/a_quick_mossimage_news_roundup_25th_october_2026/ (see this post for link to very old posts)
```



document fetched 2026-11-15 00:36 · canonical https://reddit.com/r/comfyui/comments/1xw5y0y/a_quick_mossimage_news_roundup_14th_november_2026
