Nodes/ComfyUI_MileHighStyler/Clothing_Style Styler
ComfyUI Node

Clothing_Style Styler

86 fabric patterns and clothing looks

By TripleHeadedMonkey·Created 3 years ago·Updated about a year ago· 66
Clothing_Style Styler
    • text_positive
    • text_negative
    text_positive
    text_negative
    clothing_style
    log_prompttrue

    This is the wardrobe-pattern menu, and it's one of the more workmanlike, safe-for-anything nodes in the pack. Clothing_Style is 86 entries of how fabric looks: argyle, camouflage, checkered, floral print, pinstripe, plaid/tartan, polka dot, striped, faded. If you keep typing "she's wearing a dress" and getting a plain flat-color dress, this is the menu that adds the texture and pattern you forgot to ask for.

    It pairs naturally with the other clothing nodes - Clothing_StateStylerAdvanced handles the condition of the clothing; this one handles the look of the fabric. Different axis, and you can stack them.

    How it works

    MileHighStyler is a fork of SDXL Prompt Styler. Each pattern is a JSON template with a {prompt} slot; the node drops your text_positive in and wraps the pattern phrase at a strong weight (~2.0). These are ordinary descriptive words - "plaid," "floral print" - so they carry well even where the weighting itself doesn't apply. Output is a rewritten prompt string into your CLIP Text Encode; nothing touches the model.

    The inputs and outputs that matter

    • clothing_style - the 86-entry pattern dropdown. none = off.
    • text_positive - your subject; the pattern gets welded onto it.
    • text_negative - optional negative text.

    log_prompt prints the finished prompt so you can see which pattern landed. Outputs are text_positive and text_negative strings, into a CLIP Text Encode.

    Installing it

    ComfyUI Manager: search ComfyUI_MileHighStyler, install, restart. Or:

    cd ComfyUI/custom_nodes && git clone https://github.com/TripleHeadedMonkey/ComfyUI_MileHighStyler
    

    then restart. No models, no requirements.txt, no dependencies - Python and JSON. Nodes appear under ali1234/stylers.

    Where people get burned

    • Pattern bleed. This is the classic clothing-pattern problem, and it's not the node's fault: tell a diffusion model "argyle" and it may cheerfully argyle the background too. If the pattern is leaking, lower how hard you push it or lean on a ControlNet/regional prompt to pin it to the garment.
    • Stacking on the wardrobe. Combine this with a clothing-state menu and your own outfit description and you can over-constrain the clothes into a mess. Pick the two axes that matter and stop.
    • Model era. The 2.0 weighting is SDXL/SD 1.5 syntax, silently discarded on 2026 LLM-encoded models (Flux 2, Z-Image, Qwen). The pattern words still work as plain text there - just type "plaid dress" and skip the weighting wrapper.
    • It's a convenience menu. The honest value is browsing patterns you'd forget existed, not doing anything you couldn't type. That's a fine reason to keep it; just calibrate expectations.
    Categoryali1234/stylers

    Inputs (4)

    NameTypeDefaultDescription
    text_positiveSTRING
    text_negativeSTRING
    clothing_styleCOMBO86 options: none, argyle, camouflage, checkered, floral print, pinstripe pattern, +80
    log_promptBOOLEANtrue

    Outputs (2)

    NameTypeDescription
    text_positiveSTRING
    text_negativeSTRING