ChatGPT Virtual Try-On Launches With New Favorites Feature

OpenAI is making another attempt at turning ChatGPT into a shopping companion. On Thursday, the company announced two new features rolling out globally: virtual try-on for clothing and accessories, and a Favorites function for saving products to look at later.

Neither idea is new to online retail. What matters here is the context. AI assistants are still trying to figure out where they fit in consumer shopping, and OpenAI's own track record in this area is mixed.

How virtual try-on works

The try-on feature asks users to upload a selfie or a full-body photo. ChatGPT then generates an image showing how a piece of clothing or an accessory might look on them. In shopping results, the option appears as a new Try On button.

Users are not limited to items surfaced by ChatGPT. You can upload an image of a product, such as a screenshot from a store's website, and ask the assistant to try it on for you.

Both features run on ChatGPT Images 2.5, the image model OpenAI recently launched. The company says the model produces more natural lighting and richer textures, follows editing instructions more reliably, and generates images with less delay. These are OpenAI's claims, and the source does not include independent testing to back them up.

Favorites and the Library

The second addition, Favorites, lets users save products they find to a Library inside the app. According to OpenAI, these saved items sit alongside the images created with try-on, so a user's product picks and generated looks end up in one place.

Other ways to shop

OpenAI also pointed to shopping tasks that go beyond the two new buttons. A user can describe a style they want to try and ask ChatGPT to find the pieces needed to complete the look. Another option is uploading photos of outfits worn by celebrities and asking the assistant to find matching items that are available to buy.

That second use case puts ChatGPT on ground that Pinterest and Google have held for years. Both have become go-to sources for fashion inspiration, and both can turn that browsing into sales for online retailers.

A crowded and uneven field

OpenAI's history in AI shopping is short but already includes one reversal. An earlier instant checkout feature did not perform well, and the company had to change course.

Others in the space have also run into trouble. Agentic AI startup Instinct recently began pushing product recommendations to users without being asked. Some users felt these proactive suggestions went too far and looked more like advertising than help.

Competition is a further issue. Google launched its own virtual try-on feature last year, so ChatGPT is arriving after an established rival rather than ahead of one.

Our Take

This launch looks like a more cautious play than instant checkout. Instead of trying to own the transaction, OpenAI is focusing on the earlier steps of shopping: discovery, visualisation and saving ideas. That is a lower-risk position, and it lines up with what image generation models are already good at.

It also fits a wider pattern of companies placing AI directly inside everyday consumer tasks, from AI-powered travel search to food orders placed by text message. The common question across all of them is whether people actually want to shop through a chatbot instead of a dedicated app or a search engine.

For readers, a few points are worth watching. First, whether Images 2.5 holds up to OpenAI's claims on realism, since a try-on image that looks off defeats the purpose. Second, how users respond to uploading personal photos for this, a privacy trade-off the source does not address. Third, whether ChatGPT can pull fashion discovery away from Pinterest and Google, which already have both the habit and the retailer relationships. The Instinct backlash suggests that how these features are presented matters as much as what they can do. If shopping help starts to feel like advertising, users may push back.