Kurta and kurti catalogs, shot to one look.

Western apparel is mostly about fit. Ethnic wear is about surface: chikankari, block print, bandhani, gota and mirror work. That surface is exactly what buyers zoom in on, and exactly what generic AI tools change.

• Repeating prints drift. A block print has a fixed repeat. A model that "redraws" it will shift the spacing or swap the motif.
• Embroidery gets simplified. Fine chikankari turns into a vague texture unless the detail is carried over from the reference.
• Sets have three products in one. Kurta, pant or palazzo, and dupatta each need to be visible, or the customer can't tell what's included.
• Consistency sells the range. If every kurta in a collection is on a different model under different light, the collection page looks like a marketplace, not a brand.

We treat the garment surface as fixed and only generate what's around it: the model, the pose, the setting and the light.

Most kurta brands sell the same style in four or five colours, and the same kurta alone or as a set. That multiplies images fast. A 40-style range in 4 colours with 6 shots each is 960 images.

Two things keep that manageable:

1. One look, locked early. We agree the model, backgrounds and light on three styles first. Every other style follows that look, so the collection page reads like one shoot.
2. Files mapped to SKUs. Images arrive named by SKU and shot type, so your team uploads in bulk instead of renaming by hand. If you're on Shopify, the bulk editor can write product fields and images across the catalog from one prompt.

See how the same approach works for sarees and for western workwear on model.