Announcing Drape: An SDXL Adapter for On-Model Fashion Images

Today, we release Drape, our first adapter for generating on-model images of clothing. You give it one flat-lay photo of a garment, and it generates that garment worn by a model.

What is Drape?

Drape is an adapter built on SDXL (Stable Diffusion XL). The adapter conditions each generation on your flat-lay photo, so the model starts from your product and not from a text description of it.

What does Drape take from the product photo?

• •Shape and fabric: the cut and material of the garment come from the photo, and the model renders how they fall on a body.
• •Prints and details: patterns, prints and trims carry from the product photo into the generated image.
• •Color: colors come from your product photo, not from a prompt.

Why does Drape matter for fashion brands?

On-model images usually come from a photoshoot, which ties each new product to studio, photographer and model schedules. It also limits a brand to the models it can book. With Drape, fashion brands can:

• • Make on-model images without booking a shoot for each product
• • Start on-model images from the flat-lay photo, without waiting for a shoot date
• • Show products on diverse body types and ethnicities
• • Test market response before committing to inventory

What are the limits of Drape?

Drape is a first version. Complex items, with intricate textures, seams or embellishments, can take several generations before one is right. Reducing that variability is the work we continued in Drape2, our FLUX-based successor.

How do I generate on-model images with Uwear today?

This post describes Drape as it launched. The models available today are listed on the platform. The fastest path is to ask your AI assistant through the Uwear MCP server. The API covers pure automation, and Studio is the manual path.