ChatGPT adds virtual try-on and doc-to-PDF scan

OpenAI's update positions ChatGPT as more than a text assistant, moving into visual and practical tasks. The try-on feature builds on ChatGPT Images and works on both web and mobile, with a selfie or full-body photo saved as a reference for future try-ons. This reference photo can be changed or deleted in the personalization settings. The feature is already available in Japanese ChatGPT for domestic e-commerce products, suggesting OpenAI is tailoring shopping tools to local markets.

The document scanning tool, meanwhile, simplifies a common office chore. By photographing multiple pages in succession, ChatGPT automatically combines them into one PDF ready to upload in the chat. This is initially limited to iOS, with a gradual rollout, so Android users may have to wait.

Google has been quicker to market here: it launched AI try-on in Japan in October 2025 through Google Search and Google Shopping. OpenAI's entry means shoppers now have two major platforms offering similar virtual fitting, which could put pressure on retailers to make product data compatible. How widely OpenAI's try-on is adopted may hinge on how well it handles diverse body types and lighting conditions, and whether users trust uploading personal photos to a chat app.