Free tool · no sign-up
AI-referred traffic to retail sites grew 393% year on year in early 2026, and it converts better than organic. Assistants can only recommend a store they can read, price and trust. Paste a product URL and see whether yours qualifies.
A human shopper forgives a missing delivery estimate and works it out. An agent comparing you against ten other stores simply picks one of them instead.
Can it fetch your pages at all, and is there product data and a feed to read once it gets there?
Price, currency, stock, identifier, delivery and returns — in the markup, not just on the page for a person to read.
Nothing on the buying path that only a human can clear, and payment rails an agent can plug into.
A real, identifiable business with policies and ratings it can point to when it spends someone's money.
The other half
This check looks at whether an agent could complete a purchase. The AI visibility check looks at whether an assistant can read and quote your pages in the first place. Most stores fail one and assume they've passed the other.
Shopify publishes its own readiness scanner, and several agencies and protocol groups publish checklists. Run them too — more scrutiny on your product data is a good thing.
Three things are different here. We sell you nothing on the result, so the fixes aren't shaped by what a platform wants you to adopt. We check whether you can be found as well as bought, which are separate failures most stores assume are one. And we name what we couldn't observe rather than scoring you on a guess:
Both. The rails are being built now — OpenAI and Stripe's Agentic Commerce Protocol, Google's AP2, and agent programmes from the card networks — and the merchants wired in first will be the ones agents can buy from. But nothing here depends on those standards. Complete product data, an accessible site and a checkout that doesn't challenge machines are what every one of them is built on top of, and they pay for themselves in ordinary search regardless.
Because that is where the offer lives. Run it on a homepage and we will still check your crawler access, trust signals and bot walls, but we cannot judge price, stock, identifiers or returns — so we mark those as not applicable instead of failing you for them.
No. We request one page, your robots.txt, a public feed path and one well-known path. We never add to a basket, never submit a form and never touch checkout. Anything that could only be confirmed by buying is reported as unknown.
Run both — a second opinion on your product data costs nothing. The difference is what each is for. A platform's scanner exists partly to show what that platform does well, and it grades you on the stack you are already on. This one is independent, it pairs buyability with whether an assistant can read and cite your pages at all, and it tells you plainly what it could not see rather than implying the check was exhaustive.
Partly, and it is the best position to be in. The platform can supply the checkout and the rails; it cannot invent the product data you haven't entered. Missing GTINs, stock status, delivery and returns are merchant-side, and they are the checks most stores lose points on.
No. The check runs on request, returns the result and keeps nothing. No sign-up and no database.
We run a growth engine that tests, learns and compounds — on our own sites first, then on yours.
Talk to us