OpenAI brings virtual try-on and product favorites to ChatGPT shopping

ChatGPT gains virtual try-on and Favorites

OpenAI has launched two shopping features globally in ChatGPT: a virtual try-on tool for clothing and accessories, and a Favorites function for saving products to a personal library. According to TechCrunch, the company announced the updates on Thursday, marking its latest attempt to turn the assistant into a practical shopping companion.

How the new features work

The try-on capability lets users upload a selfie or a full-body photo and see how a garment or accessory would look on them. A new Try On button will appear inside ChatGPT's shopping results, TechCrunch reports. Users can also work in the other direction: upload an image of an item, such as a screenshot from a website, and ask the assistant to show how it fits on them.

Favorites addresses a different part of the shopping flow. It lets users save products they discover into a Library within the app so items can be revisited later. TechCrunch notes that saved products sit alongside a user's try-on images.

Built on a new image model

OpenAI says both features draw on the newly released ChatGPT Images 2.5 model. As reported by TechCrunch, the company claims the model produces more natural lighting and richer textures, follows editing instructions more reliably, and generates images with less delay — qualities that matter when the output needs to be a plausible photo of a person wearing specific clothes.

ChatGPT can also assist with shopping in broader ways. TechCrunch describes scenarios such as describing a desired style and asking the assistant to find the pieces needed to complete the look, or uploading photos of a celebrity's outfit and asking it to locate purchasable versions of the items being worn.

A crowded field and a mixed track record

The launch lands amid experimentation, and some retrenchment, in AI-assisted commerce. TechCrunch reports that OpenAI previously abandoned an instant checkout feature after it failed to gain traction. Meanwhile, agentic AI startup Instinct has been pushing product recommendations to users, a move some people perceived as overreaching and closer to advertising than helpful suggestion.

ChatGPT's fashion discovery ambitions also place it in direct competition with Pinterest and Google, which in recent years have dominated style inspiration and discovery that converts into sales for online retailers. TechCrunch points out that Google introduced its own virtual try-on capability a year earlier, and it remains an open question whether users will treat ChatGPT as their first stop for this kind of activity.

Why it matters

The addition of try-on and Favorites shows AI assistants evolving from answering questions into mediating the full shopping journey: discovery, evaluation, visualisation and collection. If users adopt these tools, ChatGPT could capture a share of the product discovery traffic that currently flows to search engines and social platforms, which is why retailers and advertisers will be watching closely. The context also matters. OpenAI's own checkout experiment fell flat, and rivals' proactive product recommendations have already drawn criticism for feeling like ads. The line between genuinely useful shopping assistance and intrusive commerce is thin, and the reception of these features will test whether people actually want an AI assistant in the fitting room.