India Attire

If you're going to try exactly one node from ComfyUI Globetrotter Nodes, make it this one. The India Attire node is clearly the pack's showpiece: it's one of only two countries (with Indonesia) that ships full regional data, and it has more body-part dropdowns than any other node in the pack - headwear, forehead marks, ear jewelry, nose rings, mehndi on hands and palms, anklets, waist chains, chest threads, the works. Want a woman in a Kanjeevaram saree with a bindi, jhumkas and mehndi, standing at the Taj Mahal at golden hour? This node writes that whole paragraph for you.

What it is

A prompt-generator node for Indian characters, built from JSON data in the pack's data/ directory. Pick a person and a wardrobe, get one prose prompt STRING out, wire it into CLIP Text Encode as your positive prompt.

How it works

The mechanism is the pack standard, turned up to eleven. At load time the node reads regional appearance data (data/appearance/in.json - Punjab, South India, East, West, North East, Central), cultural context (data/cultural/in.json - Holi, Diwali-adjacent festivals like Durga Puja, Onam, Ganesh Chaturthi, playing tabla, practicing yoga), poses, and attire files spread across every body-part category. On run it assembles the prompt in priority order, gender-filters the attire, and expands each garment into a full description - a saree isn't just "saree," it's the drape, the fabric, the blouse. That level of description is exactly what makes or breaks a cultural prompt.

The inputs that matter

• region - one of the few places in the pack it genuinely matters: Punjab, South India, North East India, etc. Each region carries a physical description (skin tone, hair) that lands in the subject line, so a Punjabi subject and a South Indian subject don't come out looking identical.
• Upper_Body - the deepest list: Kanjeevaram Saree, Lehenga Choli, Anarkali Suit, Kurta, Choli, Punjabi Patiala Suit, and more.
• Forehead, Ears, Nose, Hands/Palms - Bindi, Tilak, Sindoor, Jhumka, Kundal, Nath, mehndi in multiple styles. This is where the Indian node leaves every other country node behind: nobody else in the pack covers face and hand jewelry.
• Legs/Foot/Ankles - Saree, Salwar, Churidar, Jutti, Kolhapuri Chappal, Payal (Anklet).
• cultural_element + setting - Holi Festival, Practicing Yoga, Weaving Textiles, with settings like Taj Mahal Backdrop and Kerala Backwaters. The scene-building actually works here.
• randomization + seed - off/light for consistent batches, full for exploration (full ignores the seed by design). The author's tooltips on both are worth reading.
• lora_trigger - put your character LoRA's trigger word here and it lands in the subject line. custom_prompt appends anything. experimental_llm_rewrite needs transformers; it downloads distilgpt2 on first use and can get repetitive, so treat it as optional garnish.

Output

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

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 for the LLM rewrite feature

Restart ComfyUI, or search "Globetrotter" in ComfyUI Manager. No model downloads - all Indian data ships as JSON.

Troubleshooting

• Overwhelmed by options: start with gender, region, Upper_Body, and one or two jewelry dropdowns, everything else on "none." Add detail once the basics work.
• Attire disappears when you switch gender: that's the gender filter, not data loss - the lists are male/female/unisex tagged.
• Output varies run to run: randomization defaults to moderate. Set it to "off" with a fixed seed for reproducible characters.
• experimental_llm_rewrite no-ops without transformers installed; it fails silently by design.

This is the node to reach for when you want a culturally specific Indian character and you don't want to hand-write a paragraph of saree-and-jewelry description yourself.