Who, how and why

An independent product
with documented limits.

Private Try-On is an independent product created and operated by Yan Gao, an independent creator based in China. It helps people see separate clothing pieces together before buying. It creates an AI style preview; it is not a fitting room, size recommendation or product-authentication service.

Who is behind the product?

Yan Gao designs, tests and operates the product as an independent creator. Private Try-On is not presented as an incorporated company, certified fashion company or laboratory. This page does not claim a professional credential, customer count or performance result that has not been verified.

How are previews made?

You provide a person reference and one or more clothing images. The service sends the required inputs through its server to OpenAI (GPT Image), with Agnes as backup, then returns the generated result. The model can change a face, pose, background, print, logo or garment detail. Based on written confirmation from Agnes support, inputs sent to Agnes for generation are not used for model training. We cannot link a public Agnes policy for this. This statement applies only to Agnes; it does not establish training practices at fal. Still images are generated by OpenAI (GPT Image), with Agnes as backup. OpenAI states that API inputs and outputs are not used to train its models by default. Runway video sends the generated still to fal (H3 Max Turbo), or to Agnes when selected or enabled as a fallback. Providers may retain inputs after generation; fal retention has not been verified here and immediate deletion is not claimed.

Why does it exist?

A top, pair of trousers and shoes can each look appealing in separate product photos but still feel wrong together. Private Try-On is designed to make that combination easier to judge before checkout. The purpose is a clearer styling decision—not a promise that an item will fit, drape or look identical in real life.

Product facts

What evidence is available?

We publish the inputs, unchanged result, method, observations and limitations from a documented separate-item test. It is one first-hand example, not an accuracy rate or a customer testimonial. Review the documented full-outfit test.

Where should I go for help?

Last reviewed September 21, 2026. AI style preview only. Not a guarantee of fabric, fit or sizing.

See how the product works →