Germany Attire

Base models are surprisingly bad at traditional German dress. A dirndl comes out looking like a generic sundress, and lederhosen often render as inexplicable leather shorts with no structure. The Germany Attire node from ComfyUI Globetrotter Nodes fixes the prompt side of that: it names the actual garments - Dirndl, Lederhosen, Trachtenhemd, Janker, Haferlschuhe - and attaches a real description to each so the model has something to work with.

What it is

One of the country nodes in the manifestations/comfyui-globetrotter pack. It's a prompt generator, not an image node: you pick a person and some garments, it returns one prose prompt STRING, and you wire that into your positive prompt via CLIP Text Encode.

How it works

Standard pack mechanics. The node is generated at load time from JSON in data/, assembles the prompt in priority order (subject → appearance → emotion → attire → pose → setting → atmosphere → composition → quality), filters attire by gender, and expands each garment name into a descriptive phrase. Germany's data is on the thinner side - no regional appearance data, no country-specific poses or landmarks - but the traditional garments it does carry are the ones people actually ask for, especially around Oktoberfest season.

The inputs that matter

• Upper_Body - Dirndl, Trachtenhemd, Janker. The Dirndl is the one most people want; the description field does the heavy lifting on the bodice and apron details.
• Legs - Lederhosen. Feet - Haferlschuhe (the classic Bavarian leather shoes).
• Head - Tirolerhut (Tyrolean Hat), Bollenhut (the black hat with the big pompom balls), Gamsbart.
• gender and age set the subject; the gender filter decides which garments appear in the dropdowns.
• randomization (off/light/moderate/full) and seed control reproducibility - off + fixed seed for consistent characters, full for exploration (and it deliberately ignores the seed).
• lora_trigger, custom_prompt, experimental_llm_rewrite - the usual tail. LLM rewrite needs transformers and pulls distilgpt2 on first run.

Output

One prompt STRING → CLIP Text Encode (Prompt) → KSampler. Standard.

Install

cd /path/to/ComfyUI/custom_nodes
git clone https://github.com/manifestations/comfyui-globetrotter
cd comfyui-globetrotter
pip install -r requirements.txt   # optional - only needed for LLM rewrite

Restart ComfyUI, or find it in ComfyUI Manager by searching "Globetrotter." No model downloads.

Troubleshooting

• Want more than four garment categories? This node only ships Head / Upper_Body / Legs / Feet. The pack is data-driven and designed for extension - drop your items into data/attire/*/de.json following the schema in the README and restart.
• Output varies between runs: randomization defaults to moderate. Set it to "off" and keep the seed for reproducible results.
• Garments missing for a gender: the male/female/unisex tags filter the lists; that's intended behavior.

It's a narrow node - traditional Bavarian and alpine wear, plus a little general Austrian-alpine flavor - but for that specific brief it's genuinely the fastest way to a correct-looking prompt.