Adjective Styler (Advanced)

The SDXL dual-CLIP version of AdjectiveStyler: a menu of ~1,100 adjectives - abandoned, ablaze, abnormal, aberrant - that welds one descriptor onto your prompt at a strong weight. Same honest caveat as the basic node applies double here: this is one of the pack's "grammatical" toys, more a way to explore unexpected descriptors than a serious styling tool, and wrapping it in per-encoder text routing doesn't change that. If you just want to try an adjective, you almost certainly want the plain node.

Where it's fun is batching. Fix a seed, cycle a few odd adjectives, and see which ones do something interesting to a scene. As a discovery aid, fine. As a load-bearing part of a pipeline, no.

How the "Advanced" plumbing works

SDXL runs two text encoders, CLIP-G and CLIP-L. This node feeds them separately - text_positive_g to G, text_positive_l to L - and applies the chosen adjective to both, wrapped at the pack's strong default weight (~2.0). MileHighStyler is a fork of the classic SDXL Prompt Styler. negative_prompt_to (Both / G only / L only) routes any built-in negative; Both by default.

The inputs and outputs that matter

• adjective - the ~1,100-entry dropdown; none is off.
• text_positive_g / text_positive_l - your subject per encoder; same text in both if unsure.
• text_negative - your negative, routed by negative_prompt_to.

log_prompt prints the assembled prompt so you can see the word in context. Six outputs: text_positive_g, text_positive_l, merged text_positive, and text_negative_g / text_negative_l / merged text_negative. Feed _g / _l into a CLIPTextEncodeSDXL, or just take the merged text_positive / text_negative into a normal 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

• One adjective, one nudge. Even at 2.0, an adjective is a light touch. For a real look, reach for the artist, camera, or big style menus.
• Abstract words make noise. Many of the ~1,100 have no clear visual referent; the model can't render them. That's the toy working as designed - A/B on a fixed seed.
• Advanced when basic would do. The G/L split only matters with a CLIPTextEncodeSDXL and different per-encoder text. Otherwise use the merged outputs, or the plain AdjectiveStyler.
• Model era. Dual-CLIP is SDXL-only; meaningless on the 2026 LLM-encoded models (Flux 2, Z-Image, Qwen). On those, just type the adjective - it works fine as literal text.