OpenAI is adding an AI-powered virtual try-on feature to ChatGPT shopping, allowing users to generate an image of themselves wearing supported clothing and accessories before making a purchase.
The feature, called Try On, appears alongside compatible apparel and accessory listings in ChatGPT’s shopping results. Users can upload a selfie or take a new photo, after which ChatGPT creates a visual simulation featuring the selected item.
Try On is not limited to products surfaced by ChatGPT’s shopping engine. Users can also upload a screenshot of a garment or accessory and ask the chatbot to create a virtual try-on image. The experience appears to rely on OpenAI’s ChatGPT Images 2.5 image-generation model, which is designed to improve texture reproduction, natural lighting and preservation of a person’s features from the original image.
A reference photo uploaded for shopping remains saved for future sessions, though it can be changed or deleted in settings. OpenAI is also introducing Favorites, letting users save products to their ChatGPT library and organize them into folders alongside images created through virtual try-on.
The additions build on OpenAI’s Shopping Research experience launched last year and bring ChatGPT into closer competition with Google Shopping, which has offered AI virtual try-on for apparel for some time.
