Environment Styler

EnvironmentStyler is the "where is this happening" menu, and it's a big one - 283 entries running from the obvious (Beach, Desert, Forest, Mountain, Urban, Studio, Indoor) out into a long tail of more specific settings. If you've got a character but no world for them to stand in, this drops a location into your prompt so you're not stuck against a gray void or the model's default backdrop. It's a scene-setting shortcut, and with 283 options it doubles as a "surprise me" button when you want an unexpected backdrop.

Of all the pack's menus, this is one of the easiest to justify: environment is a huge part of an image, and it's the thing beginners most often forget to prompt at all.

How it works

MileHighStyler is a fork of SDXL Prompt Styler. Each environment is a JSON template with a {prompt} slot; the node drops your text_positive in and wraps the location phrase at a strong weight (~2.0) so the setting reads clearly. Location words are concrete nouns the model knows well, so they land reliably. Output is a rewritten prompt string into your CLIP Text Encode; nothing touches the model.

The inputs and outputs that matter

• environment - the 283-entry location dropdown. No option (or none) is the off switch.
• text_positive - your subject; the setting attaches to it.
• text_negative - optional negative text.

log_prompt prints the finished prompt to the console so you can see which location landed - handy on a menu this deep. 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

• A strong environment can swallow the subject. At 2.0, "Beach" doesn't just add a beach - it can turn the whole image into a beach with your character as an afterthought. If the setting is dominating, that's the emphasis; a lighter touch or a second styler to reassert the subject helps.
• Environment vs. composition tug-of-war. Pair a busy environment with a tight portrait cue and they fight over the frame. Decide whether you want a wide establishing shot or a close-up, and let one win.
• 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 location word still works there as plain text - just name the place in a sentence.
• 283 entries, uneven quality. A menu this large has some very specific, thinly-trained locations that the model renders vaguely. If one does nothing, it's just underrepresented - pick a more common setting.