OpenAI Launches Virtual Try-On and Product Collections in ChatGPT
OpenAI has introduced new shopping capabilities to ChatGPT globally, allowing users to virtually try on clothes and bookmark items into collections. The features are powered by the updated ChatGPT Images 2.5 model, which delivers more realistic lighting, richer textures, and faster image generation. This update pushes ChatGPT further from a text-based assistant into a visual e-commerce discovery hub. By integrating generative computer vision directly into shopping workflows, OpenAI is positioning ChatGPT as a competitor to visual search and digital retail platforms. Users can upload a selfie or full-body photo to see how apparel looks on them via a new "Try On" button in search results, or upload celebrity images to search for matching products. Saved items and generated try-on previews are stored together in the user's personal in-app library.
## BACKGROUND
Virtual try-on (VTO) technology leverages computer vision and image editing models to accurately align garments onto human figures while preserving pose and lighting details. Multimodal AI models allow users to combine text instructions with image inputs to modify generated visual content seamlessly.