Research · 6 min read
Can ChatGPT see your Shopify store? We checked 1,200 stores
Almost always, yes. In a random sample of 1,200 Shopify stores, 1 of 1,184 stores open to Google blocked ChatGPT's search crawler, and Shopify served AI agents its own instruction file on 96%. The gaps are in the product data: only 37% of product pages put a GTIN in their markup.
Key takeaways
- In our November 2026 crawl of 1,200 randomly sampled Shopify stores, 1 of 1,184 stores open to Google blocked OAI-SearchBot, the crawler behind ChatGPT search. Five more stores were closed to every bot.
- Only 51 stores named GPTBot in robots.txt, and 5 blocked it from product pages. Blocking GPTBot keeps content out of OpenAI's model training, not out of ChatGPT search.
- Shopify serves AI agents an llms.txt with shopping instructions on its own: 1,107 of 1,156 stores had it, and all 1,156 answered with a Universal Commerce Protocol profile.
- In product pages from 582 of the stores, 92% had Product markup and 84% showed the price without JavaScript, but only 37% carried a GTIN and 11% a return policy in the markup.
- Shopify answered 53 of our 66 product page requests with 'too many requests', even at one page every 15 seconds, while robots.txt, sitemaps and llms.txt always loaded. Its own llms.txt asks AI agents to identify themselves with Web Bot Auth.
A store owner who asks whether ChatGPT can see their store usually means three things. Is the store open to OpenAI’s crawler? Does an AI agent find the products and their details? And is there anything to switch on? We checked all three on a random sample of Shopify stores, with the same rules our free AI visibility tools use.
Can ChatGPT see a Shopify store?
Almost always. Of 1,189 stores in our sample that served a robots.txt, 1,183 let OAI-SearchBot, the crawler behind ChatGPT search, read their product pages. Of the six that didn’t, five were closed to every bot, Google included. One store blocked ChatGPT search while staying open to Google.
Shopify also talks to AI agents directly. 1,107 of 1,156 stores served an llms.txt file that Shopify wrote for them: instructions for AI agents on how to find products and shop the store. All 1,156 answered with a Universal Commerce Protocol profile, the machine-readable description of the store’s shopping services.
Data table
| Check | Stores or pages | Of | Percent |
|---|---|---|---|
| What any crawler can read: robots.txt lets OAI-SearchBot read product pages | 1183 | 1189 | 99.5% |
| What any crawler can read: robots.txt lets GPTBot read product pages (training crawler) | 1179 | 1189 | 99.2% |
| What any crawler can read: llms.txt with agent instructions from Shopify | 1107 | 1156 | 95.8% |
| What any crawler can read: Universal Commerce Protocol profile (/.well-known/ucp) | 1156 | 1156 | 100.0% |
Do Shopify stores block GPTBot and other AI crawlers?
Rarely, and almost never the ones that matter for search. 1,118 of the 1,189 robots.txt files in our sample named no AI crawler at all; the AI rules we found were added on top of Shopify’s file by a merchant or an app. 71 stores named at least one AI crawler. 51 named GPTBot, and 5 of them blocked it from product pages.
| Crawler | Used for | Stores blocking product pages |
|---|---|---|
| OAI-SearchBot | ChatGPT search | 6 of 1,189 (5 closed to every bot) |
| ChatGPT-User | pages a ChatGPT user asks to open | 5 |
| GPTBot | OpenAI model training | 10 |
| PerplexityBot | Perplexity search | 6 |
| Claude-SearchBot | Claude search | 5 |
| ClaudeBot | Anthropic model training | 7 |
| Google-Extended | Google AI apps and model training | 8 |
| Applebot-Extended | Apple model training | 7 |
| Meta-ExternalAgent | Meta AI training | 9 |
| Amazonbot | Amazon services | 14 |
| CCBot | Common Crawl, an open dataset used for training | 14 |
| Googlebot | Google Search, AI Overviews and AI Mode | 5 (all closed to every bot) |
Blocking the training crawlers is a separate decision from search. OpenAI says that “disallowing GPTBot indicates a site’s content should not be used in training generative AI foundation models”, while “sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers”, though they can still appear as navigational links. Google says Google-Extended “does not impact a site’s inclusion in Google Search”.
What does Shopify show AI agents by default?
A file written for them, and a protocol to shop through. The llms.txt that Shopify serves on a store starts with “Agent Instructions” and points agents to two routes: the Shop skill, and the store’s Universal Commerce Protocol (UCP) service with an MCP endpoint for finding products, building carts and checking out. It lists the store’s policies and asks agents to place orders only with the buyer’s authorization.
The same file sets rules for agent traffic. “Agents must identify their requests as agent-initiated,” it says, and server-to-server requests “must use Web Bot Auth”, a way to sign requests so a store can recognize the agent. It also tells agents to “back off on 429 responses and respect Retry-After when provided.”
Shopify’s route into ChatGPT doesn’t depend on crawling at all. In March 2026 Shopify announced that “millions of merchants can sell to ChatGPT users via Agentic Storefronts”, and its help center says agentic storefronts are “active by default for eligible stores”, with products sent through Shopify Catalog. Shopify also notes that blocking AI crawlers in robots.txt “affects only open-web discoverability” and “doesn’t stop your product data from being sent by Shopify Catalog to the agentic storefronts that you’ve activated.”
Only 47 of 1,156 stores replaced Shopify’s file with their own llms.txt, and 2 had none. You don’t need an app to get an llms.txt on Shopify.
Can AI crawlers read your product page without JavaScript?
On most pages, yes: the product name, price and Product markup are in the HTML. In product pages from 582 stores, 83% showed the product name and 84% the price in the page text, and none was an empty JavaScript shell. That matters because, in Vercel and MERJ’s 2024 analysis, “none of the major AI crawlers currently render JavaScript”, with Google’s and Apple’s crawlers as the exceptions. What an AI crawler gets is the raw HTML.
Data table
| Check | Stores or pages | Of | Percent |
|---|---|---|---|
| Product page, read without JavaScript: Product structured data | 533 | 582 | 91.6% |
| Product page, read without JavaScript: Price in the page text (pages with a price in the markup) | 440 | 521 | 84.5% |
| Product page, read without JavaScript: Product name in the page text | 482 | 582 | 82.8% |
| Product page, read without JavaScript: Brand in the markup | 533 | 582 | 91.6% |
| Product page, read without JavaScript: Availability in the markup | 530 | 582 | 91.1% |
| Product page, read without JavaScript: Barcode in the page data (from the Shopify product record) | 267 | 582 | 45.9% |
| Product page, read without JavaScript: GTIN in the markup | 217 | 582 | 37.3% |
| Product page, read without JavaScript: Rating in the markup | 100 | 582 | 17.2% |
| Product page, read without JavaScript: Shipping details in the markup | 57 | 582 | 9.8% |
| Product page, read without JavaScript: Return policy in the markup | 66 | 582 | 11.3% |
The gaps are in the details that let an agent match and compare products. A GTIN identifies the exact product across stores, and Google lists it among the recommended properties for merchant listings, along with brand, availability, ratings, shipping details and the return policy. In our sample, 46% of pages carried a barcode in the product data Shopify puts on the page, but only 37% put a GTIN in the markup: on 82 pages, the barcode was there and the markup left it out. Ratings were in the markup on 17% of pages, shipping details on 10% and the return policy on 11%.
Google says there are “no additional requirements to appear in AI Overviews or AI Mode” and no special structured data to add. Structured data still has to match what the page shows, and it is the cleanest way to hand a crawler the facts a shopper asks about.
Why did Shopify turn our crawler away from product pages?
Because we didn’t identify ourselves the way Shopify asks agents to. Shopify answered 53 of our 66 product page requests with status 429, “too many requests”. It made no difference whether we sent two requests a second or one every 15 seconds, or whether we waited as long as each Retry-After asked. Meanwhile robots.txt, sitemaps, llms.txt and UCP profiles loaded for every store that had them, at three requests a second across all stores.
We read this as Shopify throttling bots it can’t verify, which is what its llms.txt describes. Real AI crawlers come from published addresses and can sign their requests, so they are likely treated differently. For a store owner, the lesson is about measurement: a third-party “AI visibility” crawler may see far less of your store than OAI-SearchBot does, and a failed check from such a tool is not proof that ChatGPT can’t read the page.
How to get your products on ChatGPT
Start with the checks that a merchant controls:
| Step | What to check | Where |
|---|---|---|
| Keep search crawlers allowed | OAI-SearchBot, PerplexityBot, Claude-SearchBot, Googlebot and Bingbot can read /products/ | our AI bot access checker |
| Decide on training crawlers separately | blocking GPTBot or ClaudeBot doesn’t remove you from AI search | our AI robots.txt generator |
| Check the agentic channels | in Shopify admin, Sales channels › Agentic: which AI channels are on | Shopify Help Center |
| Complete the product data | GTIN or barcode, brand, availability, shipping and return policies | Shopify product and policy settings |
| Put the facts in the HTML | name, price, stock and key specs visible without JavaScript | our What AI sees tool |
| Add full Product markup | offers, GTIN, brand, rating, shipping details, return policy | our product schema generator |
Shoppers and agents also compare what the product looks like in use. How many stores show it in video, from the same crawl, is in our census of product videos on Shopify. If you want product-accurate video on your own pages, we start with a pilot on 3–5 products.
Sources
- Quelta Shopify store crawl, November 8, 2026: robots.txt, llms.txt, UCP and product page checks per store (CSV)
- OpenAI, Overview of OpenAI crawlers
- Shopify News, agentic commerce momentum (March 24, 2026)
- Shopify Help Center, Agentic storefronts
- Shopify Help Center, Editing robots.txt.liquid
- Google Search Central, Google's common crawlers (Google-Extended)
- Vercel and MERJ, The rise of the AI crawler (December 17, 2024)
- Google Search Central, AI features and your website
- Google Search Central, Merchant listing (Product, Offer) structured data
- Common Crawl, October 2026 crawl archive (CC-MAIN-2026-43)
- Tranco, a research-oriented top sites ranking (list 94P72)
- Shopify Help Center, connecting a third-party domain (A record 23.227.38.65)