Apply SDXL Style MRE

Apply SDXL Style MRE is the "high-budget action movie" member of the mozman pack. Its 22 styles (mre-*) skew dramatic, cinematic, and dark-fantasy: mre-cinematic-dynamic, mre-dark-cyberpunk, mre-heroic-fantasy, mre-undead-art, mre-dark-dream, mre-space-art, mre-sumi-e-symbolic, mre-manga. These aren't neutral presets - they're tuned for hero shots and moody concept work, and the positives read like it: "epic cinematic shot of dynamic {prompt} in motion. main subject of high budget action movie. raw photo, motion blur. best quality, high resolution."

If you're doing character key art, fantasy illustration, or anything where "dramatic" is the vibe, this set lands faster than most of what you'd write by hand. The negatives are also decently built, which matters more on SDXL than people give it credit for.

One honest note before you go looking: nobody reliably documents where this set came from. "MRE" is unexplained, and a sweep of community discussion turns up nothing on it - it circulates as sdxl_styles_mre.json, a community collection that got bundled into this pack. That's fine for a prompt template; just know you're trusting the set's tuning, not an official source.

How it works

Same machinery as every node in this pack: pick a style, the node drops your prompt into the template's {prompt} placeholder and appends the template's negative to yours. Strings in, strings out, before CLIP Text Encode. No model, no API, no downloads.

Inputs and outputs that matter

• positive_prompt / negative_prompt - your prompt strings (multiline).
• style - 23 entries: bypass plus 22 mre-* styles. bypass passes everything through unchanged.
• apply_negative_style (default yes) - flip off if you're supplying your own negative or running a model where negatives do nothing.
• log_prompt (default off) - console before/after; handy for a set this un-documented.

Outputs are positive_prompt and negative_prompt (STRING) → your CLIP Text Encode node.

Install

Part of ComfyUI_mozman_nodes - one install gets all seven nodes. ComfyUI Manager (search ComfyUI_mozman_nodes), or:

cd ComfyUI/custom_nodes
git clone https://github.com/mozman/ComfyUI_mozman_nodes

Restart after. No requirements.txt, no model files, pure stdlib, MIT-licensed, from Manfred Moitzi (the ezdxf author). Nothing here can break your environment.

Where people get burned

The obvious one: these styles are aggressively cinematic, so they're a poor fit for clean product shots or subtle work - "epic cinematic shot of dynamic" as a wrapper is going to drag a lot of drama into the frame. And as with every node in the pack, the templates assume an SDXL-lineage CLIP encoder; on guidance-distilled or LLM-encoded models the negative half is inert at CFG 1 and much of the phrasing just doesn't transfer.

Last practical tip: with a set this thematic, log_prompt is your friend. Turn it on once, run a couple of styles, and read what actually gets injected - it's the fastest way to know which of the 22 you'll actually use.