Google’s AI try-on feature for clothing now works using only a selfie

Google announced on Thursday that it is updating its AI-powered try-on tool to let users virtually try on clothing with just a selfie. Previously, users needed to upload a full-body photo to generate a virtual outfit preview. Now, a simple selfie is enough for the system to generate a full-body digital representation using Nano Banana’s Gemini 2.5 Flash Image model.

After uploading a selfie, users select their typical clothing size, and the tool produces multiple generated images. They can then choose one to use as their default try-on model.
For those who prefer it, the original option to upload a full-body image remains available, along with the ability to pick from a diverse set of preset models representing various body types.

The updated feature begins rolling out today in the United States.

Google originally debuted its AI try-on tool in July, enabling users to preview apparel from retailers included in its Shopping Graph across Google Search, Google Shopping, and Google Images. To activate the feature, shoppers tap a product listing or apparel image and choose the “try it on” option.

This update arrives as Google continues expanding its efforts in the AI-driven virtual try-on space. The company is also developing a standalone app called Doppl, designed to help users visualise how different outfits might look on them using AI.

Earlier this week, the Doppl app received a significant update: a new shoppable discovery feed featuring tailored recommendations and AI-generated visuals of real products. Nearly all items in the feed can be purchased directly via merchant links.
The feed also provides AI-generated videos and outfit suggestions based on a user’s personal style — a format Google likely views as familiar and engaging for users accustomed to TikTok- and Instagram-style content.