See your store the way ChatGPT sees it — free
Shoppers have started asking assistants what to buy. "Best full-grain leather wallet under $200." "Which store sells hand-thrown stoneware that ships to Canada." ChatGPT, Perplexity and Gemini answer those questions by reading storefronts — and they can only recommend what they can actually read.
Most merchants have never seen what those assistants find. So we built a scanner that shows you, free: lunalink.ai/scan.
What it does
Put in your store's domain. In about thirty seconds you get a score out of 100 and a list of findings, worst first. Every finding comes with a link to the page on *your* store that proves it, because a claim about your business that you can't verify in a minute isn't worth making.
It reads only what a crawler reads:
- robots.txt — whether GPTBot, PerplexityBot, ClaudeBot and Google-Extended are allowed in at all
- Your catalogue endpoint — whether the product data is readable
- Product page markup — whether the JSON-LD carries price, availability and identity
- Policy pages — shipping and returns, the two things a shopping answer needs before it recommends anyone
No install. No login. No credentials. Nothing a shopper couldn't load themselves.
Why the outside view is the honest one
There are more thorough audits — our app runs one from inside Shopify with your session and paid crawls. But an AI assistant doesn't have your session. It sees your public surface and nothing else. So the outside view isn't a compromised version of the real check; for this question, it *is* the real check.
That framing decides what the scanner will and won't say. If it can't read your catalogue, it tells you the scan was incomplete rather than reporting a clean score — because "we saw nothing" and "nothing is wrong" are different sentences, and only one of them is true.
What it finds on real stores
Three patterns come up again and again.
Product schema missing or half-filled. A model reading a product page without JSON-LD has to guess the price from page text, and usually declines to state one at all. A store we scanned had schema on half its sampled pages, and of those, one was missing `offers.price` entirely. It will get described in an answer, but not recommended in a comparison.
No product identifiers. No GTIN, MPN or SKU in the markup means a shopping surface can't tell that your listing and another retailer's listing are the same physical product. That's precisely where purchase-intent questions end up.
Blank product types. One store had 40% of its products with no product type set. Shoppers ask AI for categories — "leather wallet" — not for model names. Those products are invisible to category matching even when the titles read beautifully to a human.
None of those are exotic. All three are fixable in an afternoon, and none of them require buying anything from us.
Then what
Fix what it found. The scanner tells you how, in plain words, for every finding — and most of it is work you can do yourself in Shopify admin and your theme editor.
If you'd rather it were done for you, that's what the studio does, and Descriva is the app we use to do it: it works from inside your store, probes ChatGPT, Perplexity, Claude and Gemini directly to see whether they actually mention you, and publishes the copy, FAQ and schema they extract from.
But start with the scan. It costs nothing, it takes thirty seconds, and you'll know where you stand before you talk to anyone — including us.