Apply a friendly, warm style to your Moss-Image illustrations.

(See the Krea 2 version or the Anima version for more examples and inspirations)

Using the Friendly Sketch Moss LoRA

Hit 'remix' on one of the example images. If you're a ComfyUI fan, grab a workflow embedded on any of these example images here; they cover both plain-SFT and Turbo LoRA runs.

This LoRA was trained on Moss-Base and I run it on Moss-SFT - per the Moss team's recommendations - but you can try it on Moss-Base also if you like.

My preferred running style at the moment is Lodestones' 8-step DMD LoRA (the version available as of Oct 21 2026) at 8 to 10 steps and 2MP size, Euler / Simple and 1.0 CFG. Try other variations, there are lots of good ways to use Moss!

Update to this advice: I'm now finding the above setup with 12 or 15 steps even better - good balance of speed and quality, works here on Civ with 1.5MP generation sizes or in your own ComfyUI at 2MP

Playing with the style

Friendly Sketch rewards play and experimentation and so does Moss-Image :)

• no trigger word

• 1.0 strength, but try 0.8 to 1.2 and you'll find cool stuff

• using your prompt to specify happiness, laughing, smiling etc. gets the best results - it's built to be friendly! But of course you can use it other ways too

• by its nature it aims to produce a fun result with some visual descriptors: "high waisted riding breeches" or "a bandeau top with thick diagonal blue and white stripes" - just enough to give it something to work with

Moss is interesting because it enables a, shall we say, certain type of output out-of-the-box without an extra LoRA or whatever. So you don't need to worry about spicier outputs destroying the style. That said, I have not yet tested this with many other LoRAs mixed in or with checkpoints that bake in alternate style info - I would love to hear from you how it's going once you try some things!

Happy sketching!