OpenAI just handed ChatGPT a new role. The AI chatbot can now generate images of clothes and accessories on a user’s own body. The feature rolled out globally on Oct. 1, powered by an updated image model that promises better lighting, textures and consistency.
Users tap a “Try On” button next to clothing in shopping results. They upload a selfie or full-body photo. ChatGPT produces a visualization. Or they grab a screenshot from any retail site and ask the bot to apply that item instead. The reference photo stays saved for future attempts. Change or delete it anytime in settings under personalization.
The Mechanics Behind the Mirror
This arrives alongside a Favorites tool. Shoppers save products to a Library inside the app. Try-on images live there too. Both draw on ChatGPT Images 2.5. OpenAI says the model delivers more natural lighting and richer textures. It preserves subjects in reference photos better. It follows editing instructions more reliably across multiple turns. Latency drops.
“Images 2.5 produces more natural lighting and richer textures, is better at preserving the subjects in your reference photos, and follows editing instructions more reliably across multiple turns,” the company told WWD. “The most meaningful images feel grounded in the real-life people, places, and memories they’re based on.”
The update builds directly on earlier shopping experiments. ChatGPT already lets users describe a desired style and receive product suggestions. Upload a celebrity outfit photo. It identifies similar buyable items. Yet previous efforts at direct checkout fell short. OpenAI shifted focus. Discovery and visualization now take center stage. (TechCrunch first reported the global launch and detailed the paired Favorites function.)
But. This isn’t the first virtual try-on. Google added a similar capability to its Search in 2025. Users upload photos there too and preview outfits. Pinterest and specialized fashion apps have offered variations for years. What sets ChatGPT apart? Its conversational core. Ask it to build a full look. Request alternatives in different colors. The chat keeps going. The images evolve with the dialogue.
Retailers watch closely. A 1.2 billion weekly active user base, as noted in WWD, means even modest adoption could shift traffic. Brands may need sharper product data and imagery optimized for AI ingestion rather than traditional web catalogs. One jacket photo. Dozens of body types and poses generated on demand. No more shooting every variant.
And yet accuracy remains a question. AI still struggles with complex patterns, unusual fits or specific fabric drape. Early demos shared by outlets look promising. Natural poses. Realistic shadows. But real-world tests will decide if shoppers trust the output enough to buy. OpenAI warns about uploaded images. By default, personal account uploads can train future models unless users opt out. The same caution that applies to every ChatGPT interaction.
Industry reaction split along familiar lines. Fashion tech veterans see an evolution of tools already in use by companies like Stitch Fix or Amazon. Others point to deeper integration potential. ChatGPT doesn’t just show the shirt. It can suggest matching pants, explain why the color flatters certain skin tones, then save the whole ensemble. The Library becomes a personal style archive.
Recent coverage highlights the timing. B&T noted on Oct. 2 that the move puts added pressure on marketers to prioritize discoverability inside AI interfaces. Traditional product pages may matter less if conversations surface items first. Partnerships like the earlier L’Oréal makeup try-on inside ChatGPT hinted at this direction.
So the fitting room goes digital in a new way. Not a static app. A persistent conversational companion that remembers past tries and builds on them. OpenAI didn’t invent virtual try-on. It placed the technology inside the most widely used AI chat platform on the planet. That distinction could prove decisive.
Executives at apparel companies now face a choice. Optimize for search engines and social feeds. Or prepare catalogs for AI agents and image generators that speak directly to consumers. The barrier between inspiration and purchase shrinks further. Whether it leads to fewer returns or simply more informed browsing remains to be measured.
One thing looks clear. ChatGPT’s shopping features keep expanding. From product links to instant checkout tests to this visual layer. The assistant inches closer to a true end-to-end experience. The mirror, once physical, now lives in the cloud. And it talks back.
OpenAI Puts ChatGPT in the Fitting Room With New Virtual Try-On Tool first appeared on Web and IT News.