Filter Styler

FilterStyler is the color-grade menu - 11 entries of photo-filter looks: Black and White, Sepia, Vintage, Monochrome, HDR, Vignette, Saturated, Warm, Cool, Contrast. Think of it as an Instagram-filter dropdown for your prompt. Want the whole image warmed up, or knocked into moody black-and-white, or given that faded vintage cast? Pick the filter and it welds the right words on. It's small and it does one thing, which is honestly refreshing in a pack full of 400-entry scrolls.

Filters are pure aesthetic steering - they don't change what's in the image, just the mood and color of it - so this is a nice finishing touch to stack after a subject and setting are locked.

How it works

MileHighStyler is a fork of SDXL Prompt Styler. Each filter is a JSON template with a {prompt} slot; the node drops your text_positive in and wraps the filter phrase at a strong weight (~2.0). These are all common photographic descriptors, so they steer color and tone reliably. Output is a rewritten prompt string into your CLIP Text Encode; nothing touches the model.

The inputs and outputs that matter

• filter - the 11-entry filter dropdown. No option is the off switch.
• text_positive - your subject; the filter attaches to it.
• text_negative - optional negative text.

log_prompt prints the finished prompt to the console. 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

• Prompt filters aren't a post-process. This asks the model to generate a sepia image; it doesn't apply a LUT to the output. If you need an exact, repeatable grade, do it as an actual image filter after generation. This is a suggestion baked into the prompt, and it varies run to run.
• Filters conflict. "Warm" and "Cool" are opposites for a reason - don't stack both. And a strong "Black and White" filter will fight any color you asked for elsewhere in the prompt; the filter usually wins.
• Model era. The 2.0 weighting is SDXL/SD 1.5 syntax, silently dropped on 2026 LLM-encoded models (Flux 2, Z-Image, Qwen). The filter word still works there as plain text. For serious color control on modern models, note that naming a specific film stock (Kodak Portra, Cinestill 800T) is a more reliable grade lever than a generic "vintage" - something to type by hand, since it's not in this 11-item menu.
• It's a convenience. Eleven looks you could type. Fine for browsing, no ceiling beyond that.