You sell clothes, jewelry, shoes or accessories, and today you saw that ChatGPT can now show a shopper how an item looks on them. A customer may have mentioned it, or a headline did, and the question that follows is a practical one: are my products in there, and if not, what do I have to do? The ChatGPT virtual try-on launched on October 1, 2026, and TechCrunch's report describes it as a way to preview clothing on yourself inside ChatGPT. You do not need to build a fitting-room app or buy anything to find out where your shop stands. This post tells you what the feature does, how products get into ChatGPT shopping, which product details to check first, and what to tell customers who use it. The work is tidying product data you already have, and you can start on it in ten minutes.
Key Takeaways
A "Try on" button on clothes and accessories listings turns a selfie into a generated image, and OpenAI says the image does not guarantee fit or size.
Shopify catalogs are integrated through Shopify Catalog, and Etsy sellers may be connected too. Other shops can apply to provide a product feed, and approval and placement are not guaranteed.
Stable IDs, factual titles, current price and stock, and clean photos of the real item are what feeds ask for, and you can check them yourself before paying anyone.
ChatGPT shows a Try on button on clothes and accessories listings
When ChatGPT shows a product listing for clothes or accessories, a "Try on" button can appear on it. The shopper selects it, takes or uploads a selfie, and gets a try-on image made with ChatGPT Images. Shoppers can also upload a photo of an item in a conversation and ask how it could look on them. OpenAI describes the feature in its Shopping with ChatGPT help page.
Two things in that description matter to you. First, the image is generated, so it is a preview of the look and not a measurement. OpenAI warns that it may not represent the product or the person exactly and does not guarantee fit or size. Second, shoppers are told to confirm final price, availability, size, color, shipping and return terms on the retailer's site. So, the sale still depends on your product page being right.
Google has its own virtual try-on for apparel too, which Google describes on its shopping blog. If you already keep your product data clean for Google, most of that work carries over.
Shopify and Etsy shops may already be connected
Products appear in ChatGPT shopping from merchant and product data, supplied either by third-party providers or directly by merchants. OpenAI's help page on shopping results says Shopify catalogs are integrated through Shopify Catalog, and Etsy and Shopify participation was announced earlier for commerce features. If your store runs on Shopify, OpenAI says individual merchants do not need extra work for that catalog connection. Etsy sellers may be connected without a separate application as well.
Connected does not mean guaranteed. Which products ChatGPT picks and shows is its own decision, and it is not something you can pay for. Whether a particular listing gets the Try on button depends on the product, the merchant and the account.
If you are on another platform, you can apply. The ChatGPT merchants page says merchants can apply and that a self-serve merchant platform is planned. OpenAI's commerce developer guide says product-feed onboarding is currently open to approved partners. So, applying does not promise approval, placement or try-on eligibility. Apply through OpenAI's current merchant process and confirm your eligibility with OpenAI directly.
Your product data decides how well a listing can be read
A product feed is a structured list of your products, one row for each item or variant. OpenAI's product feed specification asks for stable product or variant identifiers, concise titles, factual descriptions, product URLs, brand, price, availability and media. Recommended extras include variants, fulfillment details and reviews. Google Merchant Center asks for similar basics: product links, image links, titles, descriptions, price, availability and brand.
For clothing, the details that cause trouble are the variant ones. Color, size, audience, material, pattern and the link between a shirt and its other sizes should all be consistent. A product page, a feed and your structured data should agree on price and stock.
| Field | What it should look like | Common slip |
|---|---|---|
| Title | Short, factual, names the item | Promotional wording or all caps |
| Description | Material, cut, care, dimensions | One generic line copied across products |
| Price and availability | Match the live page and the feed | Sale price in one place, full price in another |
| Variants | One ID per size and color, grouped to the parent | Sizes listed as separate unrelated products |
| Main image | The real item on a clean background | Banner text, watermark or placeholder |
| Product URL | Opens the exact item | Link to a category page |
| Structured data | Product and Offer markup with price and stock | Markup missing or out of date |
Google's merchant listing guidance supports Product and Offer markup for price, availability, shipping and returns, and asks that the information describe the item you actually sell. Its image guidance asks for accurate product images without promotional overlays, placeholders or generic graphics.
Truthful listings and photos are a rule, not a preference
OpenAI's commerce policies say sellers must provide truthful listings and imagery, and may not misrepresent price, availability, origin, condition or key characteristics. For a clothing shop that means your photos should show the garment you ship, in the color you ship, with measurements and materials stated accurately.
That also settles a worry you may have. You do not need new photography because of try-on. OpenAI's documented workflow uses your product listing plus the shopper's selfie or uploaded photo. If your current main images show the real item clearly and carry no overlaid text, you are in better shape than a shop with a new photo shoot and a mismatched description. If a main image is a placeholder or a lifestyle shot where the garment is hard to see, replace that one first.
Keep any size chart and fit notes on the product page. A generated preview will not tell a customer whether a medium runs small, so your own sizing details do that job.
Shoppers decide what happens to their selfies
When a shopper uses try-on, they upload their own photo to ChatGPT. OpenAI says reference photos can be saved for future try-ons, and that shoppers can change or delete them in Settings under Personalization, then Reference photos. Its data controls page explains that users can turn off "Improve the model for everyone," and that Temporary Chats are not used to improve models and may be kept for up to 30 days for safety. There is also a privacy overview.
For you, the practical rules are short:
- Tell customers who ask that the photo goes to ChatGPT, and that they manage it in their own ChatGPT settings.
- Describe try-on as an image preview and not a sizing tool.
- Send customers to OpenAI's current privacy and data pages instead of promising how their photos are handled.
- Do not collect or keep customer photos yourself unless you have a separate, clearly stated reason and their consent.
A two-sentence note on your product or FAQ page covers all of this, and it keeps you from making promises about a system you do not run.
A ten-minute check shows where your shop stands
You can run this without a developer, and it tells you whether you have a platform connection, a data problem, or nothing to fix yet.
Ask ChatGPT for your kind of product
Open ChatGPT and ask for something in your category, such as a black linen shirt under a price you sell at. Note whether any stores, brands or listings appear and whether yours is among them. Do not read a single result as a verdict; ask in a few different ways.
Find out how your store connects
Check whether your store runs on Shopify or Etsy. If it does, confirm your catalog is synced and active. If it runs on something else, read the merchants page and note whether an application is open to you.
Open three product pages
Choose three representative products, ideally one with several variants. Check the title, description, brand, price, availability, color, size, product URL, return policy and main image. Compare each with what a shopper sees on the live page. Then test the page on a phone.
Fix what you find in order
Start with stale prices and stock, then broken or unsuitable images, then inconsistent variants, then thin descriptions, then missing structured data. This order comes from how ordinary store catalogs break, not from a ranking of try-on failures, so adjust it to your shop. For a small catalog, most of these are a day or two of cleanup rather than a project.
You can stop worrying about a 3D body model or a rebuild
Try-on does not require a special 3D model of your garments, a new app, or a rebuilt catalog. It does not require a custom try-on tool either. The requirements that matter are accurate product data, usable images, current availability and a connection through a platform or an approved feed.
Also set aside the idea that you can pay for placement. OpenAI selects products on its own, and no vendor can guarantee that ChatGPT will show a given product or turn on try-on for it. A Google ranking does not guarantee ChatGPT visibility either, and joining a product feed does not automatically turn on Instant Checkout. OpenAI treats product discovery, merchant feeds, shopping research and checkout integrations as separate things, as its Buy it in ChatGPT announcement shows.
Paid help makes sense when the catalog is the problem
You can do the first checks yourself, and the free route should come first. Paying for help makes sense when the catalog is too big to inspect by hand, variants are inconsistent, a feed will not validate or stay in sync, images cannot be reached by crawlers, or a store platform needs development work. The right service is proportional to the data and integration problem. It is not a promise of preferred ChatGPT placement, and it should leave you with a clear list of what changed.
If you want to read more around this topic, our post on AI shopping agents and your online store covers AI agents that shop for customers, and the holiday shopping forecast post looks at the season ahead.
Is your shop ready for ChatGPT try-on?
Pick an answer to begin.
1. What does the ChatGPT Try on button give a shopper?
2. Your store runs on Shopify. What does OpenAI say about the catalog connection?
3. Which product detail should you fix first on a clothing listing?
Frequently Asked Questions About chatgpt virtual try on
Can ChatGPT virtually try on clothes?
Yes. Product listings for clothes and accessories can include a Try on button. The shopper takes or uploads a selfie, and ChatGPT Images generates a preview.
Does ChatGPT try-on show the right size?
No. OpenAI says generated images may not represent the product or appearance exactly and do not guarantee fit or size. Shoppers are told to confirm size, color, price, availability, shipping and returns on the retailer's site.
How can my store appear in ChatGPT shopping?
Use an eligible platform connection, such as Shopify Catalog, or apply to provide a product feed through OpenAI's merchant process. Applying does not guarantee approval or placement.
Do I need to do anything if I sell on Shopify?
OpenAI says Shopify product data is integrated through Shopify Catalog, so individual merchants do not need additional work for that catalog connection. Keep your product data accurate anyway.
What product data does a feed need?
Core fields are product or variant IDs, titles, descriptions, product URLs, brand, price, availability and media. For clothing, add consistent color, size, material and variant details.
Are shoppers' try-on photos private?
OpenAI says reference photos can be saved for future try-ons and shoppers can change or delete them in Settings. Point customers to OpenAI's current privacy and data settings, and do not keep customer photos yourself without a clear reason and consent.
Moving Forward
ChatGPT's try-on is a generated preview attached to product listings. For a small shop, the question that matters is whether your products are connected and whether the data behind them is correct. If you are on Shopify or Etsy, you may be connected already. If you are not, you can apply, with no promise of approval.
Your next step is small. Ask ChatGPT for products in your category, open three product pages, and fix stale prices, unusable images and inconsistent variants first. Add a short note for customers about how try-on photos work and send them to OpenAI's settings. Skip the 3D models and the promises of placement.
If your catalog is large or messy and you want a hand, Web Leveling can clean up product data, variants and structured data with you, and our custom e-commerce development work covers the platform side. We work with small and medium businesses across the country and overseas, and you can contact Web Leveling about your product catalog whenever you are ready.
Terms
Words used in this post
Tap a term to see what it means.
Virtual try-on. A feature that generates an image of an item on the shopper from a selfie or uploaded photo.
ChatGPT Images. The OpenAI image model that creates the try-on preview.
Product feed. A structured list of your products with IDs, titles, prices, availability, images and links.
Shopify Catalog. The Shopify product catalog that OpenAI says is integrated with ChatGPT shopping.
Variant. One version of a product, such as a single size and color.
Structured data. Code on a product page, such as Product and Offer markup, that states price, stock and other facts for search systems.
