Just upload a selfie, and in seconds, you can generate a personalized full-body digital avatar and try on your favorite clothes—the Google announced a key upgrade to its AI virtual fitting feature on Thursday. In the past, users had to provide a full-body photo to use this feature; now, with the Nano Banana image model from Gemini 2.5 Flash, the system can intelligently "fill in" the user's full-body outline from a regular selfie and generate realistic virtual fitting effects.
The new feature allows users to input their usual size, and the AI will generate multiple images of different postures for trying on, from which users can select the most satisfactory one as their default virtual fitting image. Of course, if you trust real photos more, you can still choose to upload a full-body photo, or directly use the platform's diverse virtual models—covering different body types, skin tones, and shapes—to meet a broader range of fitting needs.
This upgrade is now officially available in the United States. Since its initial launch last July, Google's virtual fitting feature has been integrated into Google Search, Google Shopping, and Google Images. Users just need to click the "Try On" button in the product list to intuitively preview how the clothing would look on them. This ability to generate a full-body image from a selfie significantly lowers the barrier to entry, making virtual fitting go from "something you need to prepare" to "something you can do on the spot."
This is not the only move Google has made in the AI fitting field. The company also runs an independent app called Doppl, specifically for AI-driven outfit visualization. This week, Doppl launched an updated "Shopable Discovery Feed"—a video recommendation page generated by AI, which provides outfit suggestions based on users' style preferences. Almost every item shown in the videos can be directly linked to purchase, connecting to real merchants.
