Is Your Shopify Store Blocking ChatGPT? We Checked 166 Stores
In our 17 August 2026 survey of 166 Shopify stores, none had a robots.txt rule that blocked the tested product path for the survey's citation-relevant crawlers. Five blocked only training crawlers. That is a result for this sample and test, not proof that every Shopify store is accessible to every AI assistant.
What did we check?
robots.txt is a public file that tells automated website readers which paths they may request. A rule telling a training bot to stay out is different from a rule blocking the bot used to discover pages for search.
We tried 274 domains discovered through 20 long-tail search queries. Two were bad seed entries. Of the 272 reachable domains, 13 could not be checked and 93 were not Shopify stores. The measured Shopify sample was 166.
The test evaluated whether a crawler could fetch the example product path /products/example-product under each site's rules. It did not mark a store blocked just because the file contained a Disallow line. Collection paths were excluded because ordinary duplicate-page restrictions had produced false alarms in an earlier pass.
| Measured result | Stores | Share of 166 |
|---|---|---|
| No tested crawler block | 161 | 97.0% |
| Training-only blocks | 5 | 3.0% |
| Citation-relevant product-path blocks | 0 | 0.0% |
A fetch failure was an unknown result, not evidence that the store blocked AI. The 13 failures included denied requests, server failures, an oversized response and a response-header error. They are excluded from the 166-store denominator.
Blocking training is different from blocking ChatGPT search
OpenAI distinguishes GPTBot, used for possible model training, from OAI-SearchBot, used for ChatGPT search discovery. Its guidance allows those choices to be made separately. Blocking GPTBot alone does not establish that OAI-SearchBot is blocked.
Google AI Overview belongs to Google Search. Google says Googlebot controls access for its search AI features; Google-Extended serves a different purpose in other Google systems. The survey treated Google-Extended as uncertain impact and did not count it alone as a citation block.
GEO (generative engine optimization) and AEO (answer engine optimization) both concern visibility in AI answers. A useful access check follows the particular crawler and page path, rather than treating every rule with an AI-related name as a lost recommendation.
Shopify's newer default showed up in most of the sample
We classified 135 files, or 81.3%, as Shopify's newer agent-oriented default, 29 as the older default and two as unknown. The newer pattern opened with Allow: / and linked instructions and discovery endpoints for shopping agents.
This split describes what the survey saw on 17 August. It does not show that a default file causes recommendations or that all current stores use the same version. Shopify's current help explains that its default file works for most stores and preserves rules for administrative and duplicate pages.
How to check your own store without changing it
- Open your primary domain followed by
/robots.txt. Read the live file, including any custom groups, rather than relying on an old screenshot or an app warning. - Check the crawler that matters for the surface you are investigating. For ChatGPT search, use OpenAI's current OAI-SearchBot guidance. Keep the decision about training bots separate.
- Check an actual public product page's path against the applicable rules. Do not treat ordinary cart, account or filtered-collection restrictions as a blanket product block.
- If the rules permit the path but access still fails, investigate the failed request separately. This survey did not test every network restriction, page indexing state or provider response.
Keep a copy of custom rules and follow Shopify's guidance if a specific change is needed. Do not replace the whole file with Allow: / or remove every restriction just because a tool labels it an AI problem.
Shopify also distinguishes direct website discovery from products shared through Shopify Catalog to activated agentic storefronts. A robots.txt rule is not a complete control over those catalog connections.
What this does and does not tell a merchant
For this measured sample, product-path blocks were not a common explanation for missing AI citations. That makes checking the actual answer and sources a useful next step before buying a crawler fix. Use our ChatGPT checking guide to save a baseline.
The sample was discovered through search results, not randomly drawn from all Shopify stores. It covers one example product path on one date, not every URL, custom network rule or indexing state. Zero observed blocks does not mean zero possible blocks, and permitted access does not guarantee a recommendation.
Frequently asked questions
How many Shopify stores did the survey measure?
It measured 166 Shopify stores on 17 August 2026. The team tried 274 domains, excluded two bad entries and 93 non-Shopify sites, and could not check 13 others. Failed checks are not counted as accessible stores.
Does blocking GPTBot stop ChatGPT search from citing my store?
Blocking GPTBot alone does not establish that search access is blocked. OpenAI identifies GPTBot with model training and OAI-SearchBot with search discovery, and says their controls are independent. Check the applicable search crawler and product path.
Should I delete my Shopify robots.txt file?
Do not remove rules just because an AI warning appears. Read the live file, check the relevant crawler and product path, and investigate a specific problem. Shopify's default rules serve a purpose, and custom changes should preserve needed restrictions.
Does an accessible store automatically get recommended?
No. The survey tested permission for an example product path under robots.txt rules. It did not measure a guaranteed recommendation, complete network access or every page's search eligibility.
Does zero blocked stores prove that Shopify never blocks AI?
No. Zero citation-relevant product-path blocks were observed among 166 measured stores on one date. The sample was not random, 13 domains could not be checked, and other paths or access controls can behave differently.
Sources and scope
Survey: 17 August 2026, anonymized aggregate results from AEO Tracker's robots crawler report and saved summary. The report and raw summary both say 166 measured Shopify stores. No sampled domain, store, raw file or merchant identity is published here. Current guidance checked 5 October 2026: OpenAI crawler guidance, Google AI search guidance and Shopify robots.txt guidance.