◑ Clothing

There's no clothing model here, no CLIP vision, no generation happening inside this node at all. IllustriousClothing is a curated dropdown menu that hands you Danbooru clothing tags you'd otherwise have to remember or look up - which, for Illustrious XL, is most of the battle. Illustrious was trained on Danbooru's tag taxonomy, and clothing is one of the deepest categories in it: a specific outfit tag like school uniform or wedding dress is something the model already knows cold, with zero LoRA required. This node just makes picking the right tag faster than typing it and hoping you spelled it the way Danbooru does.

How it's built

Five independent enum dropdowns, each defaulting to - (meaning "skip this one"): Outfits (111 choices - full presets like tuxedo, maid, miko), Top (77 choices), Bottoms (88 choices), General Style (35 choices - broader aesthetics like streetwear, gothic lolita, formal wear), and Headwear (54 choices). Pick as many or as few as you want; whichever aren't - get combined into one output string. There's no cross-checking between them, so picking both a full Outfits preset and a separate Top/Bottoms combo will just stack all of it into the prompt - that's on you to avoid if you want a clean result.

Two booleans control formatting: Inject Clothing (default on) folds the selections into the node's prefix content rather than only living in the node's own output, which matters if you're chaining several EasyIllustrious selector nodes together into one prompt. Format Tag wraps output using \( and \) - useful because a fair number of Danbooru tags contain literal parentheses (character-disambiguation tags are the classic example), and escaping them keeps CLIP's weighting syntax from misreading a tag as an emphasis group.

prefix and suffix (optional strings) let you splice existing prompt text around whatever this node generates, which is the pattern this whole node family uses to chain together without a separate "combine strings" node.

The one output

PROMPT - a plain string. Feed it into your IllustriousCLIPTextEncoder (or the stock CLIP Text Encode) alongside whatever else is building your positive prompt, or chain it into the next selector node's prefix input.

Installing it

Open ComfyUI Manager, search "EasyIllustrious," click install, restart ComfyUI - that's the path the author recommends. If you'd rather do it by hand: cd ComfyUI/custom_nodes && git clone https://github.com/regiellis/ComfyUI-EasyIllustrious, then pip install -r ComfyUI-EasyIllustrious/requirements.txt inside your ComfyUI environment, then restart.

Common issues

Nothing exotic here since the node is just string assembly, but a couple of things are worth knowing. First, this is a stock-Illustrious-friendly node: everything in its lists is Danbooru vocabulary, so it works cleanly on Illustrious, NoobAI, and their merges alike - there's no e621-only clothing terminology mixed in. Second, if your final image ignores half the clothing tags, that's a prompt-length problem more than a node problem: Illustrious runs on SDXL's 77-token CLIP pair, so stacking Outfits and Top and Bottoms and General Style and Headwear on top of a character, pose, and scene description will blow past what the model can attend to. Pick the outfit tags that actually matter for the shot and let the rest go.