Google has announced an update to its AI try-on feature that now allows users to virtually try on clothing using only a selfie, starting Thursday. Previously, the feature required a full-body photo, but now it leverages Google’s Gemini 2.5 Flash Image model, known as Nano Banana, to create a full-body digital version from a selfie.

Users can input their typical clothing size, and the system will produce multiple generated images. From these options, individuals can select one to set as their default try-on photo. For those who prefer, the original method of uploading a full-body picture or choosing from a diverse set of models remains available. This enhanced capability is launching today in the United States.

The try-on feature was first introduced in July, enabling users to virtually try on clothing items from Google’s Shopping Graph across Search, Google Shopping, and Google Images. To access it, users tap on a product listing or apparel result and select the “try it on” icon.

This update is part of Google’s ongoing investment in AI-powered virtual try-on technology, which includes a dedicated app called Doppl. The app is designed to help users visualize how various outfits might look on them using artificial intelligence.

Earlier this week, Google enhanced Doppl with a shoppable discovery feed that shows personalized recommendations, allowing users to explore and virtually try on new items. Nearly every item in the feed is shoppable, with direct links to retailers. The feed includes AI-generated videos of real products and suggests outfits tailored to individual style preferences.

While some may have reservations about an AI-generated feed, Google likely sees it as an effective way to present products in a format that users are already accustomed to, similar to popular platforms like TikTok and Instagram.