Google’s AI try-on feature for clothes now works with just a selfie

Google has unveiled an exciting update to its AI try-on feature, enabling users to virtually model clothing using just a selfie, a significant improvement from the previous requirement of a full-body photo. This innovative approach leverages Google's Gemini 2.5 Flash Image model, known as Nano Banana, to create a complete digital avatar of the user, simply based on a facial image. With this enhancement, users can select their typical clothing size, after which the feature generates multiple images for them to choose from. This allows for a personalized virtual fitting experience where users can make one of the generated images their default try-on photo. For those who prefer the classic method, the option to upload a full-body image or select from a variety of models representing different body types remains available. Launching today in the United States, this feature builds on the initial rollout of the try-on capability in July. Users can access this feature through product listings on Google Search, Google Shopping, and Google Images by tapping on the “try it on” icon next to apparel items. This move is part of Google’s broader investment in AI-driven virtual try-ons, which includes their dedicated app, Doppl. Doppl was designed to help users visualize how various outfits will appear on them through AI technology. Recently, Doppl received an update that introduced a shoppable discovery feed, which showcases AI-generated videos of real products and offers outfit suggestions tailored to the user's style. Almost all items in this feed are linked directly to merchants for easy purchase. While some users might be hesitant about an AI-curated feed, Google seems to be embracing this trend, drawing inspiration from popular platforms like TikTok and Instagram to present products in a familiar and engaging format.