Can ChatGPT see your store?
Robots rules, structured data, llms.txt, content shape — scored out of 100 in one scan.
Enter a valid store domain — like yourstore.com.
Free · No email · Reads only your public pages
Shoppers now start with a question, not a search.
A growing share of shopping starts as a question to ChatGPT, Perplexity or Google's AI Overviews — and those engines only recommend stores they can read, parse and trust. This scan measures exactly that: whether AI crawlers are allowed in by your robots.txt, whether your structured data actually parses, whether your content is shaped to be quoted, and whether your metadata holds together. Every check maps to a concrete fix, and the whole thing runs from your public pages — no login, no access to your admin.
Straight answers.
How does ChatGPT find and recommend products?
Through crawlers — OpenAI's GPTBot and OAI-SearchBot, Anthropic's ClaudeBot, PerplexityBot and others — that read the public web, then through models that quote what they could parse and verify. If the crawlers are blocked or your data is unreadable, engines either skip you or describe you second-hand.
Does blocking GPTBot matter?
Blocking AI crawlers is opting out of being read and cited. Some publishers do it deliberately; for a store it usually just means your products can't be recommended when a shopper asks an AI for suggestions. The scan reads your actual robots.txt rules with a real parser — group precedence, wildcards and all — and shows the verdict per bot with the exact rule that decided it.
What is llms.txt?
A convention: a Markdown file at your site root summarising what your site is and linking key pages, written for AI systems. No engine enforces it yet, but it's cheap and gives crawlers a clean map. We publish one ourselves — and the Navaal SEO app generates one for Shopify stores.
Why does structured data decide citations?
JSON-LD lets a machine read your prices, availability and answers without guessing — and engines cite what they can verify. One syntax error silently invalidates a whole block, which is why this scan parses every block individually and reports the character position of anything broken.
What can the scan not see?
It reads your public HTML the way a crawler does. Schema injected only by JavaScript after page load can't be verified externally — the scan says "can't verify" rather than guessing. And it measures whether engines can read you, not how they rank you once they do.