Prompt Composer List - Clothing

Typing out clothing descriptions gets tedious fast, and it's easy to phrase the same garment three different ways across a session without noticing. This node swaps the typing box for two dropdowns of pre-written clothing phrases - one broad, one specific - so you're picking consistent, already-worded prompt text instead of guessing at phrasing.

The Clothing in the class name is a hint that this is one instance of a category-specific "list folders" picker rather than a one-off; whatever other categories the pack ships alongside it, this is the clothing one.

How it works

Two independent dropdowns, each backed by a bundled list rather than a file you edit yourself (that's what Custom Lists is for - see that node if you want your own editable lists). Generic covers broad garment types - dress, sweater, jacket, cardigan, blouse, t-shirt, shirt, tank top, hoodie, sweatshirt - about 11 choices in total. Complex goes specific, and its entries are already phrased ready to drop into a prompt: "wearing square neck," "wearing flounce short sleeve," "wearing shirred ruffle hem," "wearing cropped turtleneck," and roughly 40 choices in that vein. Pick one from either (or both), set a weight, and the node builds the same (phrase:weight) string every other node in this pack builds.

The inputs and outputs that matter

• Generic (enum, ~11 choices, default -) - broad garment nouns.
• Generic (weight) (0–1.95, step 0.1, default 1) - weight for the Generic pick.
• Complex (enum, ~40 choices, default -) - specific, pre-phrased clothing descriptors.
• Complex (weight) (0–1.95, step 0.1, default 1) - weight for the Complex pick.
• active (boolean, defaults on) - bypass switch.
• text_in_opt (optional string) - chain input.
• text_out - the combined result.

Where it fits

Leave both dropdowns at - and it contributes nothing, so it's safe to drop into a graph before you've decided on an outfit. Wire it into a chain the same way as any other Prompt Composer node - its text_out feeds the next node's text_in_opt. If the exact phrasing you want isn't in either list, that's what text_in_opt is for: feed in your own Single Text node's output first, and this node's picks get appended to it rather than replacing it.

How to install it

ComfyUI Manager: search florestefano1975, install, restart. Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/florestefano1975/comfyui-prompt-composer

restart ComfyUI. Pure text-node pack - no model downloads, nothing heavy to wait on.

Common issues & troubleshooting

Weight isn't changing anything. The (phrase:weight) syntax this node produces is CLIP-only. It works on SD1.5, SDXL, and SDXL-lineage checkpoints (Illustrious, NoobAI, Pony). On an LLM-encoded model - Flux, Z-Image, Anima - it's inert; the parentheses just become literal text in your prompt, which is worse than leaving the weight at 1.

The garment you want isn't a choice. These two dropdowns are fixed, bundled lists - you can't add to them the way you can with Custom Lists' file-driven categories. If nothing here fits, type it directly in a Single Text node and chain it in through text_in_opt instead.

Output looks empty even though you're using the node. Check both dropdowns are actually off -. Leaving them there is the intended "not contributing" state.