One Button Preset

OneButtonPrompt's main node has a lot of knobs - insanitylevel, subject, artist, imagetype, base_model - and the pack's own advice to first-time users is to not get overwhelmed by them and just play with the presets first. OneButtonPreset is that advice made into a node: one dropdown, no other tuning required, and it emits a finished prompt built from a combination of settings the author already worked out and found good.

Reach for this one when you want output from the pack's engine but don't want to learn its whole parameter space first, or when you know roughly the kind of result you want - fantasy portraits, consistent SDXL output, a specific aesthetic - without wanting to hand-tune every widget on the main node to get there.

How it works

You pick a named preset, and internally it resolves to a specific combination of the settings the main OneButtonPrompt node exposes, then generates a prompt from that combination. A few of the named presets are documented directly by the pack's own guides:

• Standard - the pack's original default behavior: wild and unpredictable, good for open-ended exploration.
• Unique People Portraits - biased toward generating photographic people.
• D&D Style Portraits - biased toward fantasy-style character portraits.
• Consistent results - draws from a large set of built-in styles for more repeatable output; the pack's own guidance is that it works best on SDXL.
• Greg mode - The Preset - leans on characters mixed with popular artists in the Greg Rutkowski vein; works best on SD1.5-family checkpoints.

The rest of the 23 choices (Cybernetic Animals, Cyberpunk Characters, Waifu's, Husbando's, Fashion Models, and more) follow the same idea - a themed combination someone already tuned.

The inputs and outputs that matter

• OneButtonPreset (required, default "Standard") - the entire point of the node. Pick the one that matches what you're after.
• base_model and prompt_enhancer (optional) - work the same as on the main node: which SD family you're targeting, and whether to run the result through the SuperPrompt-v1 expander.
• preset_prefix / preset_suffix (optional) - this node's equivalent of prompt_prefix/prompt_suffix on OneButtonPrompt: plain text appended to the finished prompt, the usual spot for LoRA trigger words.
• seed (optional) - fix it for a reproducible pick from the preset's own randomness, randomize it for variety within the theme.
• Output: prompt, a single string. Note there's no prompt_g/prompt_l split here like on the main node - if you're targeting SDXL's dual encoder specifically, feed this into a regular CLIPTextEncode or split the string yourself, rather than expecting the SDXL-specific pair of outputs.

How to install it

Same pack, same steps: ComfyUI Manager → search "One Button Prompt" → install → restart, or manually:

cd ComfyUI/custom_nodes
git clone https://github.com/AIrjen/OneButtonPrompt

then restart. No models beyond the same optional SuperPrompt-v1 dependency prompt_enhancer pulls in elsewhere in the pack.

Common issues & troubleshooting

A preset name you remember from elsewhere isn't in the list, or vice versa. The ComfyUI node's preset list is its own; don't assume perfect 1:1 parity with the pack's Automatic1111 extension, which has its own preset selector.

Can't see what settings a preset actually resolves to. There's no inspection panel for that inside this node - it's a black box that emits a finished string, not a loader for the main node's widgets. If you want full visibility into every parameter, use OneButtonPrompt directly and set them by hand.

"Consistent results" isn't looking consistent. The pack's own documentation says that preset is tuned for SDXL - check your base_model input actually matches what you're rendering with.