OpenAI Launches Virtual Try-On and Favorites in ChatGPT

OpenAI announced the global launch of two shopping features in ChatGPT on Thursday: a virtual try-on capability for clothing and accessories, and a Favorites function for saving products to a Library inside the app.

The virtual try-on feature lets users upload a selfie or a full-body photo to visualize how a garment or accessory might look on them. The option appears as a new "try on" button in ChatGPT's shopping results. Users can also upload an image of an item — a web screenshot, for example — and ask ChatGPT to try it on for them.

Favorites lets users save products they discover to a Library in the app so they can return to them later. OpenAI notes that saved items sit alongside users' try-on images.

The move marks OpenAI's latest experiment with turning its conversational assistant into a shopping tool. The stakes are commercial and competitive: AI assistants across the industry are probing consumer use cases around shopping, and the winners could redirect a slice of retail discovery and sales that platforms like Google and Pinterest currently command.

The model behind it

OpenAI said the new shopping features leverage the newly launched ChatGPT Images 2.5 model. The company claims the model produces more natural lighting and richer textures, follows editing instructions more reliably, and reduces image generation latency. Those properties matter for try-on: the feature depends on placing garments on a user's photo convincingly enough that the result reads as a plausible preview rather than a crude composite.

Shopping as an assistant use case

OpenAI's path here has not been linear. The company already had to pivot from one of its earlier ideas in this space — an instant checkout feature that ended up not performing well.

The broader market shows the same friction. Agentic AI startup Instinct more recently began pushing product recommendations to users, but some felt that the proactive recommendations were an overreach — closer to ads than helpful suggestions. OpenAI's new features are opt-in by design: users initiate the try-on with their own photos and choose what to save to Favorites.

Beyond try-on

OpenAI said ChatGPT can support shopping in additional ways. A user could describe a style they want to try, then ask the assistant to shop for the pieces needed to complete the look. Another option: upload photos of celebrities' outfits and ask ChatGPT to find the items they are wearing that are available for purchase.

That second use case puts ChatGPT in direct competition with Pinterest and Google, which have dominated fashion inspiration and discovery in recent years as sources that convert interest into sales for online retailers.

A contested field

Whether ChatGPT becomes people's first choice for this type of activity remains to be seen. Google launched virtual try-on last year, giving it a head start with an established base of shopping-related search behavior. OpenAI's advantage, if any, is that try-on now lives inside a conversational interface where users can describe a look, upload photos, refine the result, and save items without leaving the chat.

The test ahead is whether users treat a chatbot as a fitting room. OpenAI has already shown willingness to drop shopping features that underperform, as it did with instant checkout — so adoption in the coming months will likely determine whether virtual try-on becomes a fixture in ChatGPT or the company's next pivot.