What do AI shopping
agents see on your store?
ChatGPT, Claude and Perplexity now shop for your customers — through Shopify's agent endpoints, not your theme. Paste your store URL to see exactly what they receive: your catalog, product data, FAQ answers, cart and whether you even show up in AI shopping results.
Try your own product searches (optional)
How a shopper would phrase it to an assistant. Leave blank and we'll derive searches from your catalog automatically.
Asking your store's agent endpoints what an AI shopper would get…
Free · no signup · reads Shopify's public UCP/MCP agent endpoints, the Global Catalog and robots.txt — nothing is added to any cart. Bot checks by Cloudflare Turnstile. See our privacy policy.
Agents don't shop the way people do
A shopping agent never sees your homepage. It asks your store structured questions over machine endpoints and works from whatever comes back — so a storefront that converts beautifully can still be invisible to the thing doing the buying.
Your design is not the interface
Photography, layout and copy — the things you spend the most on — are not what an agent reads. It reads your catalog feed, your product schema and your agent endpoints. Those are usually whatever the theme and apps happened to emit.
Someone gets shown instead of you
Shopify's Global Catalog answers cross-merchant questions like "waterproof hiking boots under $150". If your products don't surface for the questions your customers ask, competitors fill that slot — and the scan names the ones that appeared where you didn't.
Blocking crawlers is a decision
Plenty of stores block AI crawlers without ever deciding to — a platform default, or a plugin someone installed. It's a legitimate choice, but it should be one you made on purpose, with the traffic consequences understood.
What the scan probes
Your agent endpoints
It requests your UCP discovery document to see which capabilities you declare, then calls your MCP endpoint the way an agent would — listing available tools, running a catalog search and fetching a product — and reports what actually came back rather than what was advertised.
Where you rank in the Global Catalog
Shopping questions are put to Shopify's cross-merchant catalog to see whether your products come back, in what position, and who appears alongside them. Give it your own searches, or let it derive them from your public product feed.
Crawler access and schema
Your robots.txt and llms.txt are checked against fifteen named AI crawlers. Then the homepage is read for the Organization and WebSite schema that lets an engine identify your brand, and a product page for the Product, Offer and Breadcrumb schema an agent needs to quote a price.
What this scan will not do
This runs against stores the person scanning may not own, so some things it could technically measure are deliberately off the table.
It never creates a cart
Whether an agent can transact is a fair question, and the honest way to answer it is to check whether cart and checkout tools are exposed at all. Actually creating carts on stores at scale is abuse, so the scan reports their presence and stops there.
It doesn't need a browser
Every check is a plain server-side request to a public endpoint — the same surfaces any agent can reach. Nothing is rendered, nothing is clicked, and no JavaScript from your store is executed.
Unreachable is not "absent"
If an endpoint doesn't answer at all, that's reported as unknown. A surface we failed to reach is never graded as one you failed to build — the two are different problems and only one of them is yours.