Bot Appétit the free AEO + SEO readiness check
Scan
Research

Vercel AI readiness statistics

In November 2026, we could only tell which sites run on Vercel where our scanner could read the page, so this page covers those sites and says nothing about how many turned our scanner away.

Updated 07:07 UTC · November 2026

Based on the sites Bot Appétit has scanned. Most of these sites come from lists we scan for research. Not a random sample of the web.

All sites on the State of the Web

How Vercel sites land

Every page we read gets a score out of 100, the number at the top of its report (how scoring works). Each row is a range of scores. Where too few sites score in the lowest ranges, they are joined into one row.

  • 71–80 29%
  • 0–70 38%
Show as a table
ScoreShare of sites
71–8029%
0–7038%

A platform is only detectable on a page our scanner could read, so these shares are of the Vercel sites we could read.

Based on the sites Bot Appétit has scanned. Most of these sites come from lists we scan for research. Not a random sample of the web.

Better or worse

Not enough history yet.

Based on the sites Bot Appétit has scanned. Most of these sites come from lists we scan for research. Not a random sample of the web.

AI access

  • robots.txt disallows some AI crawlers 10%
  • robots.txt disallows major AI crawlers 3%
Show as a table
What we foundShare of sites
robots.txt disallows some AI crawlers10%
robots.txt disallows major AI crawlers3%

Based on the sites Bot Appétit has scanned. Most of these sites come from lists we scan for research. Not a random sample of the web.

What Vercel sites get right and wrong

Passed more often than on the average site: “Buttons, links and fields have names” (70%, against 38% of all sites); “Reading the page is affordable for a machine” (98%, against 85% of all sites); “Core security headers are set” (23%, against 12% of all sites). Least often passed: several checks that no Vercel site we could read passed.

Every check, Vercel against all sites

Share of sites that pass each check, over the sites the check applies to. The Gap column is the difference in percentage points between the two shares; a dash means not enough sites yet.

Check Vercel sites All sites Gap (points)
Served over a secure HTTPS connection 100% >99% Level
HTTP redirects to HTTPS 99% 93% +6 points
No mixed (insecure) content 100% 99% +1 point
Core security headers are set 23% 12% +11 points
Not flagged for malware or phishing 100% 100% Level
No server version leakage 99% 91% +8 points
security.txt is published 3% 4% −1 point
Reached without a redirect chain 100% 99% +1 point
Caching or validation headers are set 100% 94% +6 points
The site settles on one host (www vs non-www) 77% 88% −11 points
robots.txt is present and not over-blocking 58% 87% −29 points
Not accidentally set to noindex 98% 99% −1 point
This page declares its canonical URL 41% 63% −22 points
Mobile viewport is set 73% 90% −17 points
Has a title tag 100% >99% Level
Title fits a search snippet 41% 51% −10 points
Has a meta description 82% 89% −7 points
Description fits a search snippet 62% 60% +2 points
The page has an H1 36% 74% −38 points
Logical heading order 42% 61% −19 points
Topic is consistent across title, H1 and description 29% 54% −25 points
Images have descriptive alt text 91% 89% +2 points
Your page's HTML isn't oversized 99% 98% +1 point
Open Graph tags for sharing 44% 37% +7 points
Twitter/X card tags 60% 70% −10 points
XML sitemap is present and referenced 44% 70% −26 points
Missing pages return a real 404 56% 88% −32 points
Images are optimised 28% 18% +10 points
Language and charset declared 98% 90% +8 points
hreflang for language variants — 99% —
Analytics is installed 30% 70% −40 points
Search Console / Bing verification 54% 77% −23 points
A favicon is set 93% 96% −3 points
No meta-refresh redirect 100% 99% +1 point
Your content survives without JavaScript 54% 93% −39 points
Your answer is in the HTML 41% 91% −50 points
Buttons, links and fields have names 70% 38% +32 points
Semantic HTML landmarks 31% 53% −22 points
Content is scannable in self-contained chunks 40% 87% −47 points
Claims come with figures, quotes or sources 49% 54% −5 points
Section headings are phrased as questions 21% 21% Level
Reading the page is affordable for a machine 98% 85% +13 points
No hidden text written for an AI model 100% >99% Level
Sitemap dates tell a machine what changed 51% 64% −13 points
The links in llms.txt resolve — 88% —
Author / publisher is clear 39% 45% −6 points
A machine-readable published / updated date 4% 12% −8 points
llms.txt is present and valid 12% 22% −10 points
Clean content for agents (Markdown) 4% 6% −2 points
Structured data is present 36% 53% −17 points
Structured data is valid 100% 97% +3 points
FAQ / Q&A schema — 44% —
Organization schema with sameAs 13% 25% −12 points
Right schema type for the page 90% 83% +7 points
Schema facts also appear in the page text — 52% —
Breadcrumb structured data — 22% —
AI crawlers aren't blocked 73% 88% −15 points
Automated visitors can reach your pages — — —
Engines are allowed to quote the page 99% 99% Level
Your page arrives quickly 97% 95% +2 points
AI bot rules are explicit and intentional 16% 17% −1 point
Content usage preferences declared 7% 5% +2 points
Resource-discovery Link headers 2% 1% +1 point
No broken links 93% 87% +6 points
Links don't point at redirects 92% 88% +4 points
Descriptive anchor text 100% 99% +1 point
Sensible internal/external link mix 67% 94% −27 points
Links are real links 95% 85% +10 points
Largest Contentful Paint (LCP) — — —
Cumulative Layout Shift (CLS) — — —
Interaction to Next Paint (INP) — — —
Page weight and render-blocking — — —
MCP Server Card 2% 1% +1 point
AI catalog 0% 1% −1 point
A2A Agent Card 1% <1% +1 point
WebMCP tools in the page 0% 2% −2 points
AI usage preferences (Content-Usage) 0% <1% Level
Licensing terms for AI (RSL) 0% <1% Level
Text-and-data-mining reservation (TDMRep) 1% <1% +1 point
Markdown version linked from the page 1% 1% Level
Content signals agree everywhere — — —
Agent Skills index 2% 1% +1 point
API catalog 1% 1% Level
OAuth authorization-server metadata 2% 3% −1 point
OAuth protected-resource metadata 2% 2% Level
OpenID Connect discovery 1% 3% −2 points
Web Bot Auth (operates a signed bot) 0% <1% Level
Universal Commerce Protocol manifest — 10% —
Auth.md agent registration 0% <1% Level
AI discovery via DNS 0% <1% Level

Based on the sites Bot Appétit has scanned. Most of these sites come from lists we scan for research. Not a random sample of the web.

See how your site compares

A free scan checks all of this on your page in seconds, with no signup.

Scan your site free

Want it handled? We do that too. No pitch, just a plain-English chat.

Book a call