Custom AI Model with Vitiligo: Preserving What Makes Your Ambassador Unique

When a fashion brand chooses an ambassador with vitiligo, they're making a statement: beauty isn't uniform, and our brand reflects that. The depigmentation patterns on their skin aren't incidental — they're the reason this person was chosen. They represent authenticity, inclusivity, and a commitment to showing real beauty in product photography.

But most AI fashion photography tools can't handle this. They homogenize skin, smooth away distinctive features, and output a generic face that could belong to any brand. Vitiligo patterns disappear. Birthmarks vanish. The very details that make your ambassador recognizable are erased — defeating the entire purpose of choosing them.

On-Model works differently. Every distinguishing physical feature is preserved faithfully in the AI-generated output, whether you're converting flat-lay garments into on-model images or swapping your ambassador into existing product photography. This guide demonstrates that with a real custom identity: a brand ambassador with vitiligo, processed through both Flat-to-Model and Model Swap pipelines.

What you'll need

• An On-Model account — sign up free if you haven't already
• 1 reference photo of your brand ambassador (clear, well-lit, front-facing — vitiligo or other distinctive features clearly visible)
• Product images: flat-lay garments for Flat-to-Model, or existing on-model photos for Model Swap

Meet Adama — our custom brand ambassador

Adama is a 21-year-old woman with vitiligo — distinctive depigmentation patterns across her face, neck, and body that make her instantly recognizable. She wasn't selected from On-Model's default catalog. She was uploaded as a custom identity — a real person chosen as a brand ambassador precisely because her look represents authenticity and beauty in diversity.

The custom identity upload process is simple: one reference photo, uploaded through the Identity section in the app. On-Model's automatic POV augmentation generates multiple angle views from that single photo, so the identity works naturally from any pose. The entire process takes about two minutes.

Flat-to-Model — from flat-lay garments to Adama on-model

This is the core test: can AI-generated product photography preserve vitiligo patterns when generating on-model images from flat-lay inputs? We uploaded two garments and generated five variations — each a different pose, all featuring Adama.

The results speak for themselves. Across all five variations, Adama's vitiligo patterns are preserved with the same consistency you'd see in a traditional photoshoot. The depigmentation on her face, neck, and visible skin areas matches her reference portrait — not approximately, but faithfully. The garments fit naturally, lighting adapts to each pose, and the skin details that define her appearance remain intact.

This is the same output quality you'd get with any Pro-tier catalog identity. The difference is that Adama is exclusively yours — no other brand can generate images with her face.

Model Swap — replacing an existing model with Adama

Flat-to-Model generates new images from scratch. Model Swap does something different: it takes an existing on-model product photo and replaces the original model with your ambassador, preserving the garment exactly as photographed. Both the clothing details and the ambassador's distinctive features need to survive the swap.

Every garment detail from the original photo — folds, shadows, fabric texture, proportions — is preserved pixel-perfect. And Adama's vitiligo patterns appear naturally on the swapped result, consistent with her identity reference. This means brands can take an entire existing product photography library and unify it under their custom ambassador without reshooting a single image.

For fashion e-commerce teams managing thousands of SKUs, this is the practical advantage: consistent brand ambassador representation across your entire catalog, generated from existing assets, with distinctive physical features preserved throughout.

Why vitiligo preservation matters for AI fashion photography

This isn't just a technical showcase. The ability to preserve distinctive skin features in AI-generated product photography has real implications for how fashion brands approach diversity and representation.

Authentic brand representation. When customers see your ambassador across your product pages, they recognize her. If the AI smooths away the vitiligo that makes Adama distinctive, the brand story falls apart. Faithful preservation means your AI-generated images are indistinguishable in spirit from a traditional photoshoot with the same person.

Inclusive e-commerce at scale. Models with vitiligo, heterochromia, birthmarks, and other distinctive features deserve the same quality of AI-generated product photography as anyone else. If AI tools can only produce homogenized skin, they're excluding the very ambassadors that brands increasingly want to represent. On-Model treats every identity — catalog or custom, conventional or distinctive — with the same fidelity.

Consistent product photography across thousands of SKUs. The same person, with the same vitiligo patterns, wearing every product in your catalog. Whether generated via Flat-to-Model or swapped into existing photography, the result is unmistakably your ambassador. That level of consistency across an entire e-commerce catalog would be prohibitively expensive with traditional photoshoots.

Get started with your own brand ambassador

Uploading a custom identity is straightforward:

1. Upload one reference photo — Navigate to Identity in the On-Model app, click Upload New Identity, and select a clear portrait of your ambassador
2. Wait for preprocessing — Automatic POV augmentation takes 1-2 minutes to generate multi-angle views from your single photo
3. Generate product images — Use your custom identity with both Flat-to-Model and Model Swap, across all poses and styles

For a complete walkthrough of identity management — including the default virtual model library, upload best practices, and multi-style generation — see the Brand Identity Guide.