Prompt Styler Mood
Calm, mysterious, chaotic — atmosphere on a dropdown
- positive_prompt_text_g
- negative_prompt_text_g
Same subject, different feeling. That's the job here. The dropdown is a list of emotional tones - Aggressive, Calm, Chaotic, Energetic, Happy, Mysterious, Relaxed, Romantic, Sad, Serene, Ethereal Tranquility, and 20 more - and the node appends the lighting, palette, and atmosphere language that carries that mood. Keep "a woman by a window," swing it from Serene to Mysterious, and the whole picture's temperature changes.
It's part of wolfden's SDXL Prompt Styler pack, and it scratches a specific itch: you often know how you want the image to feel before you know how to say it. Beginners tend to prompt the subject well and the atmosphere not at all. This node fills that gap with 31 presets.
How it works
Each mood is a JSON template - a prompt field with a {prompt} placeholder, and a negative_prompt. The node slots your positive text into the placeholder and wraps it with the mood's descriptors, returning the assembled string. Leave your negative blank to inherit the template's; type one and the two combine. It's pure string templating: no model runs inside the node, nothing downloads, and the image is entirely up to the checkpoint you send the text to.
Mood is one of the more forgiving things to prompt, because it works through broad levers - warm vs. cold light, tight vs. open framing, saturated vs. muted color - that SDXL handles well. So this node tends to land more reliably than, say, a precise composition preset.
The inputs and outputs that matter
- style - the mood. The node's whole reason to exist; 31 tones on offer.
- text_positive - your subject and scene. The mood recolors it.
- text_negative - optional; blank uses the template's default, or add your own recurring offenders.
log_prompt (No/Yes) echoes the finished prompt to the console - useful for seeing what "Ethereal Tranquility" actually expands into.
Two outputs: positive_prompt_text_g and negative_prompt_text_g. Wire them into your positive and negative CLIP Text Encode nodes and sample.
Installing it
ComfyUI Manager: search SDXL Prompt Styler (wolfden), install, restart.
Manual:
cd ComfyUI/custom_nodes
git clone https://github.com/wolfden/ComfyUi_PromptStylers
Restart ComfyUI. No pip dependencies, no model files. Find the styler nodes under a Style Prompts submenu after restarting.
Common issues
Don't contradict your own prompt. If your positive already says "bright sunny meadow," applying Sad or Mysterious gives the model two moods to average. Pick a tone that agrees with the scene, or strip the conflicting descriptors from your subject text.
none is a passthrough. The first list item, none byMood, changes nothing by design.
Moods overlap. Calm, Relaxed, and Serene are cousins; don't expect night-and-day differences between neighbors. The bigger jumps are across the emotional spectrum - Aggressive vs. Serene.
Built for SDXL. The templates lean on SDXL-style tag prompting and a working negative prompt. On Flux or Z-Image the negative does nothing at CFG 1 and the descriptors read as literal instruction, so the mood steer softens. Run it on an SDXL-family checkpoint for the intended effect.
Inputs (4)
| Name | Type | Default | Description |
|---|---|---|---|
| text_positive | STRING | — | |
| text_negative | STRING | — | |
| style | COMBO | 31 options: none byMood, Aggressive, Calm, Chaotic, Energetic, Happy, +25 | |
| log_prompt | COMBO | No | 2 options: No, Yes |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| positive_prompt_text_g | STRING | — |
| negative_prompt_text_g | STRING | — |