A free visibility check for online stores
ChatGPT, Claude and Perplexity read only the HTML your server sends — they execute no JavaScript and fill in nothing. We checked 208 German online stores: on a good quarter of them, price or availability was missing from exactly that HTML. Usually not because it loads in later, but because it is not there at all. We show you field by field what actually reaches the assistants.
We load 10 pages at most, no more than one per second, and honor your robots.txt. Takes 30–60 seconds. See an example report →
On the left, what your customers see in a browser. On the right, the same HTML as GPTBot is served it. The most common finding across 208 stores: the price is there — availability and identifiers are not.
Nordlicht Manufaktur
129,00 €
● In stock – ships in 1–2 days
Brand: —
129,00 €
Availability: —
"@type": "Product",
"offers": {
"price": "129.00",
"priceCurrency": "EUR"
/* availability missing */
}
An assistant that cannot find availability, brand and GTIN cannot identify your product with confidence — it recommends the product it can read in full.
+393 %
Adobe Analytics analyzed over a trillion visits to US retail sites: AI-referred traffic grew 393 % year over year in the first quarter of 2026, and 1,151 % in December.
Source: Adobe Analytics, "AI traffic surge — retail sites not machine-readable"
+42 %
Within twelve months the conversion rate went from roughly half the average of other channels to 42 % above it. Revenue per visit is 37 % higher.
Source: Adobe Analytics 2026
27 %
Our own measurement across 208 German stores on WooCommerce, Shopware and Shopify: on 27 % of them, price or availability was missing from the HTML as served — availability on one store in four, price on one in eight. The major crawlers execute no JavaScript and see exactly that HTML.
Source: our own measurement, 208 German stores, August 2026
0 €
Cloudflare's "Markdown for Agents" converts the HTML it is given. A price that was never in the HTML does not appear in the Markdown either. The format gets better; the missing data stays missing.
Source: Cloudflare Markdown for Agents, Feb 2026
A study of around 300,000 domains (SERanking) found no measurable effect of
llms.txt on AI citations — and that the major crawlers barely
request the file. We report it, but we do not score it. Anyone selling you
an llms.txt as the fix is selling you a placebo.
We also never impersonate GPTBot or anybody else's crawler. Whether a bot is blocked on your site we evaluate locally against your robots.txt — just as accurate, and without the deception.
No account, no credit card. The report is publicly shareable.