rubric v0.9 · scanned 2026-10-20 17:05:33Z

vercel.com

HTTP cache checker

Your HTTP cache result

The HTTP cache checker puts these checks first. The full report below lists every check, these included.

Passed Adopted HTTP caching of the homepage — The homepage sends Cache-Control "public, max-age=0, must-revalidate" and ETag, and a request with If-None-Match was answered HTTP 304 Not Modified.
Observation
```
Cache-Control: public, max-age=0, must-revalidate · ETag: W/"mj6uko8rpactyo" · Last-Modified: (none) · conditional GET with If-None-Match: HTTP 304
```
Cache-Control says how long the homepage may be reused. A validator lets a crawler ask whether it changed, and a 304 answer skips the body.

Sources: RFC 9111: HTTP Caching (IETF, 2022-06), RFC 9110, 13: conditional requests (IETF), Google: HTTP caching for crawlers (Google Search Central, 2024-12), Bing Webmaster Guidelines, §3: ETag and lastmod (Microsoft Bing)

Read today by

Googlebot sends If-None-Match and If-Modified-Since. Every HTTP cache reads Cache-Control.

Google Search

Direct

Read the check page
Fails worth 0.7 points Adopted A real 404 for a missing page — /aws-probe-3f511295eb8c, a path that does not exist, answered HTTP 200 after a redirect to https://vercel.com/login?next=%2Faws-probe-3f511295eb8c: a soft 404. Every missing URL on the site reads as a page.
Observation
```
HTTP 200 · text/html · first 16 KB read · final https://vercel.com/login?next=%2Faws-probe-3f511295eb8c
```
A URL that does not exist answers 404 or 410 (RFC 9110). A 200 there makes every mistyped link look like a page.

Sources: RFC 9110, 15.5.5: 404 Not Found (IETF), Google: HTTP status codes and soft 404s (Google Search Central), Bing Webmaster Guidelines, §9: 404 for missing pages (Microsoft Bing)

Read today by

Google reports a missing page that answers 200 as a soft 404. Every HTTP client reads the status first.

Google Search

Direct

Note

Answer a URL that matches no page with HTTP 404, or 410 for a page removed on purpose. A single-page app that serves its shell for every path can still send 404 for a route it does not know.
Verify with
```
ORIGIN=https://vercel.com
MISSING="/aws-probe-$(head -c 6 /dev/urandom | od -An -tx1 | tr -d ' \n')"
curl -sS -o /dev/null -w 'status=%{http_code} location=%{redirect_url}\n' -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' "$ORIGIN$MISSING"
```
Read the check page
Not used worth 4.3 points Adopted Structured errors for JSON requests — Not used: /aws-probe-3f511295eb8c, asked for JSON, answered HTTP 404 with a text/markdown, not a JSON error.
Observation
```
HTTP 404 · text/markdown · 271 bytes
```
An agent that asks for JSON and gets an HTML error page has to read a page meant for people. RFC 9457 defines a JSON error format.

Sources: RFC 9457: Problem Details for HTTP APIs (IETF, 2023-07), RFC 9110, 12.5.1: Accept (IETF)

Read today by

API clients and agents that send Accept: application/json. No search engine reads it.

Google Search

None
Fix
```
{
  "type": "about:blank",
  "title": "Not Found",
  "status": 404,
  "detail": "Nothing is published at this address."
}
```
Note

Send this with Content-Type: application/problem+json and the status 404 when a request for a missing path asks for JSON. Browsers can keep the HTML page (RFC 9457).
Verify with
```
ORIGIN=https://vercel.com
MISSING="/aws-probe-$(head -c 6 /dev/urandom | od -An -tx1 | tr -d ' \n')"
curl -sS -D - -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: application/json, application/problem+json;q=0.9, */*;q=0.1' "$ORIGIN$MISSING" | head -20
```
Read the check page

Read the full report Open the HTTP cache checker

71.5

out of 100

L3 · Machine-readable

Web standards 83.7 · Agent conventions 42.9

Usable by agents: No action surface found

Held at L3: L4 Usable by agents needs openapi to pass or be partial. Read the level rules

Readable: HTTP 200 at https://vercel.com/, 537 words in the HTML. 8 to fix, 2 optional choices, 3 agent conventions not used, 1 agent convention to correct, and 24 out of scope.

How this was measured

Measured once, from 1 vantage point, at https://vercel.com. The scanner reads the homepage and at most 129 URLs, runs no JavaScript, and calls no API or MCP tool it finds.

Not a ranking prediction, not AI visibility, not a certification.

We sell no consulting. No ads, no sponsored listings, and we never contact the sites that are scanned. Scanning is free today.

Defects 8

Web-standards checks that fail or are partial, and optional files served with an error. Each lists what fixing it is worth.

Fails worth 8.4 points Adopted OpenAPI definition — No OpenAPI 3.x document was found at the 5 paths the scan requested (0 advertised, 5 standard).

Observation

tried: /openapi.json 200 | /openapi.yaml 200 html | /.well-known/openapi.json 404 | /api/openapi.json 404 | /swagger.json 200 html

An OpenAPI document describes an HTTP API so a client can call it without guessing.

Sources: OpenAPI 3.1 (OpenAPI Initiative), RFC 8631 (IETF)

Read today by: API clients and code generators.
Google Search: None

Fix

{
  "openapi": "3.1.0",
  "info": {
    "title": "Vercel Inc.",
    "version": "<choose: API version>"
  },
  "servers": [
    {
      "url": "https://vercel.com"
    }
  ],
  "paths": {}
}

Note

A document at one of the 5 standard paths counts on its own. Also point at it from the homepage, a Link header or your API catalog, so a client that reads those finds it, and fill paths with the operations the API serves.

Verify with

ORIGIN=https://vercel.com
for p in /openapi.json /.well-known/openapi.json /api/openapi.json /swagger.json; do printf '%s -> ' "$p"; curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: application/json, application/linkset+json;q=0.9, */*;q=0.5' "$ORIGIN$p" | jq -c '{openapi, title: .info.title, version: .info.version, paths: (.paths | length)}' 2>/dev/null || echo "no JSON OpenAPI document"; done
printf '/openapi.yaml -> '; curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: application/yaml, application/x-yaml;q=0.9, text/yaml;q=0.9, */*;q=0.5' "$ORIGIN/openapi.yaml" | grep -m1 -E '^(openapi|swagger):' || echo "no YAML OpenAPI document"

Read the check page

Partial worth 0.9 points Adopted API catalog — A linkset is served at /.well-known/api-catalog with 2 entries, and 1 pairs an anchor with a service-desc link. No service-desc or service-doc target was observed to answer: 1 same-origin target was not requested by this scan and 1 target on another host was not followed. Not scored: the status target is on www.vercel-status.com, which this scan does not request.
Observation
```
relations: item, service-desc, service-doc, status
```
An API catalog lists a publisher's APIs in one document at a fixed address (RFC 9727).

Sources: RFC 9727 (IETF, 2025-06)

Read today by

API clients that follow RFC 9727.

Google Search

None

Note

Point service-desc at a same-origin document the scanner requests, such as /openapi.json, and make it answer 200.
Verify with
```
ORIGIN=https://vercel.com
curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: application/linkset+json, application/json;q=0.9, */*;q=0.5' "$ORIGIN/.well-known/api-catalog" | jq '.linkset[] | {anchor, "service-desc": .["service-desc"], status: .status}'
```
Read the check page
Fails worth 0.7 points Adopted A real 404 for a missing page — /aws-probe-3f511295eb8c, a path that does not exist, answered HTTP 200 after a redirect to https://vercel.com/login?next=%2Faws-probe-3f511295eb8c: a soft 404. Every missing URL on the site reads as a page.
Observation
```
HTTP 200 · text/html · first 16 KB read · final https://vercel.com/login?next=%2Faws-probe-3f511295eb8c
```
A URL that does not exist answers 404 or 410 (RFC 9110). A 200 there makes every mistyped link look like a page.

Sources: RFC 9110, 15.5.5: 404 Not Found (IETF), Google: HTTP status codes and soft 404s (Google Search Central), Bing Webmaster Guidelines, §9: 404 for missing pages (Microsoft Bing)

Read today by

Google reports a missing page that answers 200 as a soft 404. Every HTTP client reads the status first.

Google Search

Direct

Note

Answer a URL that matches no page with HTTP 404, or 410 for a page removed on purpose. A single-page app that serves its shell for every path can still send 404 for a route it does not know.
Verify with
```
ORIGIN=https://vercel.com
MISSING="/aws-probe-$(head -c 6 /dev/urandom | od -An -tx1 | tr -d ' \n')"
curl -sS -o /dev/null -w 'status=%{http_code} location=%{redirect_url}\n' -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' "$ORIGIN$MISSING"
```
Read the check page

Partial worth 0.4 points Adopted Service status and changelog feed — A status page answered at /status (HTTP 200), but the homepage links no status or changelog feed.

A status page and a changelog feed say whether the service is up and what changed.

Sources: RFC 9727 (status relation) (IETF), RFC 4287 (IETF)

Read today by: Integrators and monitoring tools.
Google Search: None

Fix

<link rel="alternate" type="application/atom+xml" title="Changelog" href="https://vercel.com<choose: path of the changelog feed>">

Note

A link only counts if it resolves: a feed that 404s is served with an error, and one whose newest entry is months old is stale. The scan reads a status page on the scanned origin, such as /status; a status page on another host is not requested, so it is reported as not measured.

Verify with

ORIGIN=https://vercel.com
curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' "$ORIGIN/" | grep -oiE '<link[^>]+(atom|rss)[^>]*>|href="[^"]*(status|changelog)[^"]*"' | sort -u
for p in /status /changelog; do printf '%s -> ' "$p"; curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' -o /dev/null -w '%{http_code} %{content_type}\n' "$ORIGIN$p"; done

Read the check page

Partial worth 0.3 points Adopted Structured contact information — Homepage markup declares 1 organization-type node with 0 contact details, while the page itself carries 0 plain-text contact signals. Organization "Vercel Inc.": 1 ContactPoint with a contactType and no reachable email or telephone.
A ContactPoint in the Organization markup makes the contact route machine-readable.

Sources: Google: Organization structured data (Google Search Central)

Read today by

Google lists contactPoint as a recommended Organization property.

Google Search

Indirect
Fix
```
{
  "contactPoint": {
    "@type": "ContactPoint",
    "contactType": "<choose: what this contact is for, such as customer service>",
    "email": "<choose: email address>"
  }
}
```
Verify with
```
HOMEPAGE=https://vercel.com/
curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' "$HOMEPAGE" | tr -d '\n' | grep -oE '"contactPoint":|"(contactType|email|telephone)":"[^"]*"' | sort -u
```
Read the check page
Partial worth 0.3 points Adopted Responsive markup — viewport content "width=device-width, initial-scale=1, maximum-scale=1" sets maximum-scale=1, which stops a reader from zooming. WCAG 1.4.4 asks that text can be enlarged to 200%.
Observation
```
viewport meta: width=device-width, initial-scale=1, maximum-scale=1
```
Google indexes the smartphone rendering of a page. A viewport that blocks zoom fails WCAG 1.4.4.

Sources: Google: mobile-first indexing (Google Search Central), WCAG 2.2, 1.4.4 (W3C)

Read today by

Browsers; Google's smartphone crawler renders the page with it. No Google feature reads the tag itself.

Google Search

Indirect
Fix
```
<meta name="viewport" content="width=device-width, initial-scale=1">
```
Note

Never set user-scalable=no or a maximum-scale below 2: they keep a reader from enlarging text to 200%.
Verify with
```
HOMEPAGE=https://vercel.com/
curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' "$HOMEPAGE" | perl -0777 -ne 'print m{<meta\b[^>]*\bname=["\x27]?viewport["\x27]?[^>]*>}i ? "viewport: $&\n" : "viewport: none\n"'
```
Read the check page
optional file, served with an error 0 points Draft /auth.md agent authentication guide — An authentication guide is served at /auth.md but it serves an HTML sign-in page instead of a markdown guide, so an agent cannot register itself.
auth.md is one vendor's convention (WorkOS) for telling an agent how to authenticate.

Sources: WorkOS auth.md (WorkOS)

Read today by

Agents built on WorkOS.

Google Search

None

Note

auth.md is one vendor's convention (WorkOS): the file is the agent-readable guide, while the machine-readable contract lives in the agent_auth block of your /.well-known/oauth-authorization-server metadata, so publish both.
Verify with
```
ORIGIN=https://vercel.com
printf '/auth.md matches='; curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/plain, text/markdown;q=0.9, */*;q=0.5' "$ORIGIN/auth.md" | grep -cE 'identity_endpoint|claim_endpoint|identity_assertion|service_auth'
```
Read the check page
optional file, served with an error 0 points Experimental Sandbox or test environment documented — A sandbox is referenced at /docs/sandbox, but it names no non-production base URL, so an agent cannot integrate safely.
A documented sandbox lets a client test calls without touching production data.

Sources: OpenAPI 3.1 (servers) (OpenAPI Initiative)

Read today by

API integrators.

Google Search

None

Note

Merge the two servers entries into the OpenAPI document you already publish rather than replacing it: OpenAPI standardises servers but says nothing about test keys, sandbox tenants, or data resets, so those can only be stated as info.description prose. A sandbox on its own host is reported as declared, not observed, because this scan requests no other host; one on the scanned host has to answer a request without credentials. The declaring source also has to say what the sandbox excludes - no live data, test data only, resets nightly - because that statement is part of what the check reads.
Verify with
```
ORIGIN=https://vercel.com
{ curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: application/json, application/linkset+json;q=0.9, */*;q=0.5' "$ORIGIN/openapi.json" | jq -r '.servers[]?.url' 2>/dev/null | grep -iE 'sandbox|staging|test'; curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' "$ORIGIN/docs/sandbox" | grep -oiE 'https://[a-z0-9.-]*(sandbox|staging)[a-z0-9./_-]*'; } | sort -u | while read -r u; do printf '%s -> ' "$u"; curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' -o /dev/null -w '%{http_code} %{content_type}\n' "$u"; done
```
Read the check page

Optional choices 3

Show 3

Decisions only you can make. None earns points or holds a level.

Not used 0 points Adopted Snippet and AI grounding directives — Not used: the homepage has no noarchive, nocache, nosnippet or max-snippet:0 directive, and no data-nosnippet attribute, so Bing applies none of the limits they set on snippets and Copilot answers.

Observation

<meta name="robots" content="index, max-image-preview:large"> · data-nosnippet: 0 elements

noarchive, nocache, nosnippet and data-nosnippet limit how Bing shows a page and how Copilot uses it. Each is the site owner's choice.

Sources: Bing Webmaster Guidelines, §10: NOARCHIVE, NOCACHE and NOSNIPPET (Microsoft Bing), Bing: robots meta tags and attributes that Bing supports (Microsoft Bing), Google: robots meta tag specifications (Google Search Central)

Read today by: Bing reads all five, and Copilot follows noarchive and nocache. Google reads nosnippet, max-snippet and data-nosnippet.
Google Search: Direct

Note

Each directive is your choice. noarchive keeps the page out of Copilot answers, nocache lets Copilot use only the URL, the title and the snippet, and nosnippet or data-nosnippet may limit the quality of what Copilot cites.

Verify with

HOMEPAGE=https://vercel.com/
curl -sSL -o /dev/null -D - -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' "$HOMEPAGE" | tr -d '\r' | grep -i '^x-robots-tag:'
curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' "$HOMEPAGE" | perl -0777 -ne 'my $h = /^(.*?)<\/head>/si ? $1 : substr($_, 0, 8192); while ($h =~ m{<meta\b[^>]*>}gi) { my $t = $&; next unless $t =~ /\bname=["\x27]?(robots|bingbot|googlebot)["\x27]?[\s>\/]/i; my $n = lc $1; my ($c) = $t =~ /\bcontent=["\x27]([^"\x27]*)/i; print "meta $n: ", ($c // ""), "\n" } my $body = $_; $body =~ s{<!--.*?-->}{}gs; $body =~ s{<(script|style|template|title|textarea|noscript)\b[^>]*>.*?</\1\s*>}{}gsi; my $n = 0; while ($body =~ m{<[a-z][a-z0-9-]*\b((?:[^>"\x27]|"[^"]*"|\x27[^\x27]*\x27)*)>}gi) { (my $t = $1) =~ s{"[^"]*"|\x27[^\x27]*\x27}{}g; $n++ if $t =~ /(?:^|\s)data-nosnippet(?=[\s=\/]|$)/i } print "data-nosnippet: $n\n"'

Read the check page

Not used 0 points Adopted Breadcrumb markup on an inner page — The inner page /contact/sales has no BreadcrumbList in its JSON-LD, Microdata, or RDFa.

Observation

https://vercel.com/contact/sales · HTTP 200 · text/html · BreadcrumbList: 0

BreadcrumbList states where an inner page sits in the site. A homepage has no trail.

Sources: Google: breadcrumb (Google Search Central), Google blog (Google Search Central, 2025-01)

Read today by: Google shows breadcrumbs in desktop results.
Google Search: Direct

Fix

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "BreadcrumbList",
  "itemListElement": [
    {
      "@type": "ListItem",
      "position": 1,
      "name": "Vercel Inc.",
      "item": "https://vercel.com/"
    },
    {
      "@type": "ListItem",
      "position": 2,
      "name": "<choose: name of this section>",
      "item": "<choose: URL of the /contact page, if the site has one>"
    },
    {
      "@type": "ListItem",
      "position": 3,
      "name": "Contact Sales",
      "item": "https://vercel.com/contact/sales"
    }
  ]
}
</script>

Note

Replace each <choose: …> with the name or the URL the site uses for that section, and drop a level that has no page. Show the same trail as links on the page.

Verify with

PAGE=https://vercel.com/contact/sales
curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' -o /tmp/aws-page.html "$PAGE"; perl -0777 -ne 'while (m{<script[^>]*application/ld\+json[^>]*>(.*?)</script>}gsi) { print "$1\n" }' /tmp/aws-page.html | jq -r '.. | objects | select((.["@type"]? | if type == "array" then .[] else . end) == "BreadcrumbList") | .itemListElement[] | [(.position | tostring), (.name // .item.name? // "NO NAME"), ((.item["@id"]? // .item) // "NO URL" | tostring)] | @tsv'; grep -c 'schema.org/BreadcrumbList' /tmp/aws-page.html

Read the check page

Present 0 points Adopted security.txt (RFC 9116) — security.txt at /.well-known/security.txt declares 2 Contacts and an Expires 2027-09-28T12:00:00.000Z in the future. Not scored: its Policy URL is on security.vercel.com, which this scan does not request.
Observation
```
Contact: https://hackerone.com/vercel Contact: mailto:responsible.disclosure@vercel.com Expires: 2027-09-28T12:00:00.000Z Preferred-Languages: en Policy: https://security.vercel.com/ Canonical: https://vercel.com/.well-known/security.txt Hiring: https://verce…
```
security.txt tells a researcher where to report a vulnerability (RFC 9116). It has no effect on search.

Sources: RFC 9116 (IETF, 2022-04)

Read today by

Security researchers and disclosure tools.

Google Search

None

Read the check page

Agent conventions 9

Show 9

The scored ones make up 30% of the score, and the unscored ones earn no points. None has an effect on Google Search today. Each scored row not in use lists what using it is worth.

Not used worth 4.3 points Adopted AI crawler policy in robots.txt — All 18 AI crawler tokens follow the * group (17 rules); none has its own group.
Observation
```
* group: 17 rules, partly disallowed, followed by 18 AI crawler tokens · robots.txt: 1 group, 17 rules
```
A crawler follows the robots.txt group that names it, otherwise the * group. A named group states the site's rules for that crawler.

Sources: RFC 9309, 2.2.1 (IETF), Google common crawlers (Google Search Central), OpenAI bots (OpenAI)

Read today by

AI crawlers that document robots.txt tokens.

Google Search

None

Note

A crawler obeys the one group that names it and skips the * group (RFC 9309, 2.2.1), so a named group repeats every * rule it should keep. This convention counts a named group that allows / for its crawler. Whether to block a crawler is your choice, and a block earns no points.
Verify with
```
ORIGIN=https://vercel.com
curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/plain, */*;q=0.5' "$ORIGIN/robots.txt" | grep -Ein '^[[:space:]]*(user-agent|allow|disallow)[[:space:]]*:'
```
Read the check page
Not used worth 4.3 points Draft Content-Signal preferences — No Content-Signal line in robots.txt or the homepage headers. Under the policy, no signal states no preference.
Observation
```
HTTP 200 · text/plain · 575 bytes
```
Content-Signal states search, ai-input and ai-train preferences in robots.txt. No crawler documents honouring it.

Sources: Cloudflare: Content Signals Policy (Cloudflare, 2025-09), IETF AIPREF (IETF)

Read today by

Cloudflare's scanner scores it. The IETF AIPREF drafts are the standards track.

Google Search

None
Fix
```
Content-Signal: search=<choose: yes or no>, ai-input=<choose: yes or no>, ai-train=<choose: yes or no>
```
Note

Add this line to the site's own User-agent: * group in /robots.txt, below its User-agent lines. Each value is the site owner's choice.
Verify with
```
ORIGIN=https://vercel.com
curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/plain, */*;q=0.5' "$ORIGIN/robots.txt" | grep -i -B3 '^content-signal:'
curl -sSL -o /dev/null -D - -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,text/plain;q=0.8,*/*;q=0.7' "$ORIGIN/" | grep -i '^content-signal:'
```
Read the check page
Served with an error worth 4.3 points Draft /llms-full.txt expanded context — /llms-full.txt answered HTTP 200 with an HTML page (2479 KB), not a text file.
Observation
```
HTTP 200 · text/html · 2479 KB
```
A convention that puts a site's text in one file. The llms.txt proposal does not define it.

Sources: llmstxt.org (llmstxt.org), Mintlify (Mintlify)

Read today by

Some documentation tools.

Google Search

None

Note

Tools that write /llms.txt can also write /llms-full.txt from the same pages: the text of each page, under its URL. Link it from /llms.txt.
Verify with
```
ORIGIN=https://vercel.com
curl -sSL -o /dev/null -w 'status=%{http_code} type=%{content_type} bytes=%{size_download}\n' -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/plain, text/markdown;q=0.9, */*;q=0.5' "$ORIGIN/llms-full.txt"
curl -sSL -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: text/plain, text/markdown;q=0.9, */*;q=0.5' "$ORIGIN/llms.txt" | grep -i 'llms-full'
```
Read the check page
Not used worth 4.3 points Adopted Structured errors for JSON requests — Not used: /aws-probe-3f511295eb8c, asked for JSON, answered HTTP 404 with a text/markdown, not a JSON error.
Observation
```
HTTP 404 · text/markdown · 271 bytes
```
An agent that asks for JSON and gets an HTML error page has to read a page meant for people. RFC 9457 defines a JSON error format.

Sources: RFC 9457: Problem Details for HTTP APIs (IETF, 2023-07), RFC 9110, 12.5.1: Accept (IETF)

Read today by

API clients and agents that send Accept: application/json. No search engine reads it.

Google Search

None
Fix
```
{
  "type": "about:blank",
  "title": "Not Found",
  "status": 404,
  "detail": "Nothing is published at this address."
}
```
Note

Send this with Content-Type: application/problem+json and the status 404 when a request for a missing path asks for JSON. Browsers can keep the HTML page (RFC 9457).
Verify with
```
ORIGIN=https://vercel.com
MISSING="/aws-probe-$(head -c 6 /dev/urandom | od -An -tx1 | tr -d ' \n')"
curl -sS -D - -A 'AgenticWebScore/0.3 (+https://agenticwebscore.com/bot)' -H 'Accept-Language: en' -H 'Accept: application/json, application/problem+json;q=0.9, */*;q=0.1' "$ORIGIN$MISSING" | head -20
```
Read the check page
Present Adopted Discovery via Link response headers — The homepage Link header names the API catalog at /.well-known/api-catalog (rel=api-catalog). It does not name /llms.txt and a markdown variant of the homepage.
Observation
```
link: </.well-known/api-catalog>; rel="api-catalog", </.well-known/ai-catalog.json>; rel="ai-catalog"; type="application/ai-catalog+json", </.well-known/agent-skills/index.json>; rel="agent-skills"; type="application/json"
```
A Link header (RFC 8288) points a client at a resource before it parses the HTML.

Sources: RFC 8288 (IETF), IANA link relations (IANA)

Read today by

Clients that follow RFC 8288 and RFC 9727 links.

Google Search

None

Read the check page
Present Draft /llms.txt published and well-structured — /llms.txt opens with H1 "Vercel", with 7 H2 link sections and 27 links, 5.1 KB. The scan requested 8 links it names, and each answered 2xx. Not requested: 1 link on another origin and 18 same-origin links (the scan samples up to 6).
Observation
```
HTTP 200 · text/plain · 5.1 KB · headings: Vercel | When to use Vercel | How agents should use Vercel | Agent setup | Documentation | REST API and developer resou…
```
A 2024 proposal for a markdown index of a site. Google Search does not use it.

Sources: llmstxt.org (llmstxt.org), Google: AI optimization guide (Google Search Central), Lighthouse llms.txt (Chrome for Developers)

Read today by

Some documentation tools and coding agents. Lighthouse audits it.

Google Search

None

Read the check page
Present Experimental Markdown content negotiation — The homepage serves text/markdown to `Accept: text/markdown, text/html;q=0.5`, with Vary: Accept, accept-encoding (HTTP 200).
Observation
```
HTTP 200 · text/markdown · Vary: Accept, accept-encoding · markdown body: yes
```
Some coding agents ask for text/markdown. Google Search does not use markdown.

Sources: RFC 7763 (IETF), Cloudflare: Markdown for Agents (Cloudflare), Google: AI optimization guide (Google Search Central)

Read today by

Claude Code, OpenCode and Cursor send Accept: text/markdown by default (Cloudflare, 2026-04).

Google Search

None

Read the check page
Present 0 points Draft Agent Skills declaration — Agent skills index at /.well-known/agent-skills/index.json lists 9 skills, each with name, type, description, url, and a sha256 digest under $schema https://schemas.agentskills.io/discovery/0.2.0/schema.json. 9 artifacts (https://github.com/vercel-labs/agent-skills/releases/download/agent-skills-063bee94c3f4df8453406c830b0a7df0f2860278/dep…, …) are on another host, which this scan does not request (declared, not observed).

An agent skills index lists skill packages for agent runtimes. It is a draft on GitHub.

Sources: Agent Skills Discovery RFC (Cloudflare)

Read today by

Agent runtimes that read the draft.

Google Search

None

Read the check page
Present 0 points Experimental ARD resource discovery manifest — ARD manifest at /.well-known/ai-catalog.json lists 6 entries, each with identifier, displayName, and type, plus exactly one of url or data.

An ARD manifest lists a site's agent resources. Lighthouse does not count a missing one against a site.

Sources: Lighthouse agentic browsing (Chrome for Developers)

Read today by

Lighthouse (ard-schema audit).

Google Search

None

Read the check page

Out of scope for this site 24

Show 24

The scan did not ask this site for these checks. Each says why. If one should apply, the error is ours, not yours.

Sites with an API or agent surface

The scan found OAuth and an HTTP API. A check that reads another kind of surface is not asked.

Out of scope 0 points Draft A2A agent card — Not scored: this check applies to a site that runs an A2A agent, and the scan found none.

An A2A agent card describes an agent the site runs. An A2A server publishes it.

Sources: A2A specification (A2A Project)

Read today by

A2A clients.

Google Search

None

Read the check page
Out of scope 0 points Draft Cryptographically signed agent card — Not scored: this check applies to a site that runs an A2A agent, and the scan found none.

An A2A card may be signed with JWS. Signing is optional in the specification.

Sources: A2A specification (A2A Project), RFC 7515 (IETF)

Read today by

A2A clients that verify cards.

Google Search

None

Read the check page
Out of scope 0 points Experimental DNS-AID discovery record — Not scored: this check applies to a site that runs an MCP server or runs an A2A agent, and the scan found none.

DNS-AID publishes agents in DNS. It is an individual Internet-Draft with no IETF standing.

Sources: draft-mozleywilliams-dnsop-dnsaid (IETF)

Read today by

Cloudflare's scanner.

Google Search

None

Read the check page
Out of scope 0 points Draft MCP server card — Not scored: this check applies to a site that runs an MCP server, and the scan found none.

An MCP server card describes a remote MCP server. It is an optional extension (SEP-2127), not core MCP.

Sources: SEP-2127 (Model Context Protocol, 2026-10), discovery.md (Model Context Protocol)

Read today by

MCP clients that implement SEP-2127; Cloudflare's scanner.

Google Search

None

Read the check page
Out of scope Adopted MCP server answers and lists its tools — Not scored: the scan found no MCP endpoint. No server card, AI catalog or homepage link names one, and /mcp answered HTTP 200, which does not signal an MCP server.

An agent given the address of an MCP server can call its tools. The scan opens one session to see that the server answers and what it offers.

Sources: MCP 2025-11-25: lifecycle (Model Context Protocol, 2025-11), MCP 2025-11-25: transports (Model Context Protocol, 2025-11), MCP 2025-11-25: tools (Model Context Protocol, 2025-11), Agentic AI Foundation: MCP joins the Linux Foundation (Linux Foundation, 2025-12)

Read today by

MCP clients: agent apps, coding agents, and chat products that connect to MCP servers.

Google Search

None

Read the check page
Out of scope 0 points Draft MPP discovery in OpenAPI (x-payment-info) — Not scored: no OpenAPI document was found, so x-payment-info is not asked of this site.

An operation with x-payment-info tells an agent that the call is paid, and states the intent, the payment method, and the amount. MPP is a draft, and charging agents is a business decision.

Sources: MPP (mpp.dev), MPP: multi-method discovery (mpp.dev), draft-ryan-httpauth-payment (IETF (individual draft)), isitagentready.com: MPP skill (isitagentready.com)

Read today by

Agents that pay with MPP; isitagentready.com.

Google Search

None

Read the check page
Out of scope 0 points Experimental Rate limits and quotas documented — Not scored: this check applies to API platform sites; this site was measured as SaaS.

Declared limits let a client pace its calls. RateLimit headers are an IETF draft.

Sources: draft-ietf-httpapi-ratelimit-headers (IETF)

Read today by

API clients.

Google Search

None

Read the check page
Out of scope 0 points Draft Web Bot Auth key directory — Not scored: this check applies to a site that runs an A2A agent, and the scan found none.

The operator of a signing agent publishes its keys here. A site that only serves content needs none.

Sources: draft-ietf-webbotauth-httpsig-protocol (IETF, 2026-09), Cloudflare: Web Bot Auth (Cloudflare)

Read today by

Sites and CDNs that verify signed agents.

Google Search

None

Read the check page

Sites with a commercial signal

The homepage shows no commercial signal, so commerce checks are not asked of this site.

Out of scope 0 points Draft ACP discovery (/.well-known/acp.json) — Not scored: the homepage shows no commercial signal, so commerce checks are not asked of this site.

An ACP discovery document at /.well-known/acp.json tells an agent where a business's checkout API is and which services it offers. The RFC behind it is a proposal.

Sources: ACP discovery RFC (status: Proposal) (Agentic Commerce Protocol), isitagentready.com: ACP skill (isitagentready.com)

Read today by

Agents that implement ACP checkout; isitagentready.com.

Google Search

None

Read the check page
Out of scope 0 points Draft Agent purchase policy — Not scored: the homepage shows no commercial signal, so commerce checks are not asked of this site.

Whether agents may buy, and with what approval, is a policy the site states.

Sources: ACP (Agentic Commerce Protocol)

Read today by

Agent checkouts.

Google Search

None

Read the check page
Out of scope Adopted Inventory and price freshness — Not scored: the homepage shows no commercial signal, so commerce checks are not asked of this site.

availability says whether an item is in stock. A past priceValidUntil can hide a product snippet.

Sources: Google: product snippet (Google Search Central)

Read today by

Google product snippets.

Google Search

Direct

Read the check page
Out of scope 0 points Draft Agent payment protocol (UCP profile or x402) — Not scored: the homepage shows no commercial signal, so commerce checks are not asked of this site.

Agent payment protocols are early. Accepting one is a business decision. This check reads UCP and x402. ACP and MPP have checks of their own.

Sources: UCP (Universal Commerce Protocol), x402 (x402)

Read today by

Agent checkouts that implement UCP or x402.

Google Search

None

Read the check page
Out of scope Adopted Product/catalog schema — Not scored: the homepage shows no commercial signal, so commerce checks are not asked of this site.

Product and Offer markup on product pages feeds merchant listings.

Sources: Google: merchant listing (Google Search Central), Google: product (Google Search Central)

Read today by

Google merchant listings and shopping agents.

Google Search

Direct

Read the check page
Out of scope Adopted Return policy machine-readable — Not scored: the homepage shows no commercial signal, so commerce checks are not asked of this site.

A MerchantReturnPolicy states return terms in markup, also under the Organization.

Sources: Google: return policy (Google Search Central)

Read today by

Google merchant listings.

Google Search

Direct

Read the check page
Out of scope Adopted Seller identity declared — Not scored: the homepage shows no commercial signal, so commerce checks are not asked of this site.

On a marketplace, the seller named on an offer is who the buyer contracts with.

Sources: Google: merchant listing (Google Search Central)

Read today by

Google merchant listings.

Google Search

Direct

Read the check page
Out of scope Adopted Shipping cost and delivery declared — Not scored: the homepage shows no commercial signal, so commerce checks are not asked of this site.

shippingDetails lets a client compute the delivered price.

Sources: Google: merchant listing (Google Search Central)

Read today by

Google merchant listings.

Google Search

Direct

Read the check page
Out of scope 0 points Draft UCP manifest (/.well-known/ucp) — Not scored: the homepage shows no commercial signal, so commerce checks are not asked of this site.

A UCP profile at /.well-known/ucp declares commerce capabilities. UCP is a draft.

Sources: UCP (Universal Commerce Protocol), Shopify profiles (Shopify)

Read today by

UCP agents; Shopify stores serve it.

Google Search

None

Read the check page

Every site

Out of scope 0 points Experimental AI usage policy page — Not scored: robots.txt and Content-Signal state no AI restriction, so there is none for a policy page to explain.

A page that explains in prose the AI restrictions the site states. No standard defines it.

Sources: RFC 9309 (IETF)

Read today by

People reading the site's terms.

Google Search

None

Read the check page
Out of scope 0 points Draft Machine-readable booking action — Not scored: this check applies to services sites; this site was measured as SaaS.

A ReserveAction declares a booking target in markup. No search engine documents a feature that reads it.

Sources: schema.org/ReserveAction (Schema.org)

Read today by

None documented.

Google Search

None

Read the check page
Out of scope 0 points Adopted Documents in machine-readable formats — Not scored: this check applies to institution sites; this site was measured as SaaS.

Tagged PDF or an HTML version can be read by screen readers and by machines.

Sources: GOV.UK: HTML not PDF (Government Digital Service), ISO 14289 (PDF/UA) (ISO)

Read today by

Screen readers; GOV.UK publishes HTML first.

Google Search

None

Read the check page
Out of scope 0 points Adopted FAQ markup for visible questions — Not scored: no heading on the homepage names a FAQ section, and the homepage carries no FAQPage markup.
Observation
```
visible questions under a FAQ heading: 0
```
FAQPage marks up questions that are visible on the page. Google stopped showing FAQ rich results on 7 May 2026.

Sources: Google Search updates (Google Search Central), Google: structured data policies (Google Search Central)

Read today by

Bing names FAQ markup as an example of labelled content.

Google Search

None

Read the check page
Out of scope 0 points Draft Reuse and licensing terms declared — Not scored: this check applies to publisher or institution sites; this site was measured as SaaS.

TDMRep reserves text and data mining rights in a machine-readable form, as EU DSM Article 4(3) allows.

Sources: TDMRep CG report (W3C), Directive (EU) 2019/790 (European Union)

Read today by

Text and data mining tools that implement TDMRep.

Google Search

None

Read the check page
Out of scope 0 points Adopted Content provenance declared — Not scored: this check applies to publisher or institution sites; this site was measured as SaaS.

Article markup with an author and dates lets a client attribute and date a source.

Sources: Google: Article structured data (Google Search Central)

Read today by

Google article features.

Google Search

Direct

Read the check page
Out of scope Experimental WebMCP browser tool declaration — Not scored: the homepage has no form and calls no registerTool, so it offers no task to declare.

WebMCP lets a page register tools for an in-browser agent. It is a W3C Community Group draft.

Sources: WebMCP draft (W3C Web Machine Learning Community Group), Chrome: WebMCP (Chrome for Developers)

Read today by

Chrome, behind an origin trial. Lighthouse audits it.

Google Search

None

Read the check page

Passed 16

Show 16

Scored checks that pass, the heaviest first.

Passed Adopted JSON-LD structured data — 3 JSON-LD blocks parse, and the page entity, Organization "Vercel Inc.", is stated under a schema.org @context.
Observation
```
JSON-LD: 3 blocks, 3 parsed · schema.org @context: yes · Microdata: 0 items · RDFa: 0 items · top-level types: Organization, Service, SoftwareApplication
```
Structured data names the entity a page describes. Google reads JSON-LD, Microdata and RDFa alike.

Sources: Google: intro to structured data (Google Search Central), schema.org (Schema.org), Bing Webmaster Guidelines, §14: structured data (Microsoft Bing)

Read today by

Google rich results; any schema.org reader.

Google Search

Direct

Read the check page
Passed Adopted Content that needs no JavaScript — 3617 characters of text are in the HTML before any script runs.
Observation
```
static text: 3617 characters · inside <noscript>: 0 characters · markup: 598704 characters
```
Several AI crawlers do not run JavaScript. Text that appears only after scripts run is invisible to them.

Sources: Vercel: the rise of the AI crawler (Vercel), Google: JavaScript SEO basics (Google Search Central), Bing Webmaster Guidelines, §8: rendering (Microsoft Bing)

Read today by

GPTBot and ClaudeBot fetch HTML without rendering; Googlebot renders.

Google Search

Indirect

Read the check page
Passed Adopted Readable homepage — HTTP 200 text/html, 585 KB, 537 words of visible text.
Observation
```
HTTP 200 · text/html · 585 KB
```
The homepage is the first URL every agent requests. An error or an empty page gives it nothing to read.

Sources: RFC 9110 (IETF), Google: 301 redirects (Google Search Central), Bing Webmaster Guidelines, §7: 301 and 302 redirects (Microsoft Bing)

Read today by

Every crawler and agent.

Google Search

Direct

Read the check page
Passed Adopted OAuth discovery documents — RFC 8414 metadata at /.well-known/openid-configuration declares issuer https://vercel.com, authorization_endpoint, token_endpoint, and 1 response type.

OAuth metadata tells a client where and how to authorise against the API.

Sources: RFC 8414 (IETF), RFC 9728 (IETF)

Read today by

OAuth clients; MCP clients use RFC 9728.

Google Search

None

Read the check page
Passed Adopted Semantic heading structure — 1 H1, 17 H2s, and 0 H3s in an outline of 18 headings, with no skipped level.
Observation
```
outline: H1 H2 H2 H2 H2 H2 H2 H2 H2 H2 H2 H2 H2 H2 H2 H2 H2 H2
```
Headings outline the page for screen readers and for any client that splits text by section.

Sources: W3C WAI headings tutorial (W3C), Google SEO starter guide (Google Search Central), Bing Webmaster Guidelines, §13: titles, descriptions and headings (Microsoft Bing)

Read today by

Screen readers; clients that chunk HTML by heading.

Google Search

Indirect

Read the check page
Passed Adopted Homepage indexable: no noindex directive — No noindex: the homepage states <meta name="robots" content="index, max-image-preview:large">, which does not ask to leave the page out.
Observation
```
<meta name="robots" content="index, max-image-preview:large">
```
A noindex in a meta tag or in a response header keeps the page out of search results. Bing keeps it out of Copilot too.

Sources: Bing Webmaster Guidelines, §10: NOINDEX and the other robots directives (Microsoft Bing), Bing: robots meta tags and attributes that Bing supports (Microsoft Bing), Google: block indexing with noindex (Google Search Central), Google: robots meta tag specifications (Google Search Central)

Read today by

Google and Bing. Bing applies the same directives to Copilot answers.

Google Search

Direct

Read the check page
Passed Adopted Title and description — Title "Agentic Infrastructure - Vercel" (31 characters) and a description of 45 characters.
Observation
```
title (31 characters): Agentic Infrastructure - Vercel · description (45 characters): The autonomous stack for every app and agent. · first H1: Agentic Infrastructure
```
The title and the description say what a page is about. Google uses them for the title link and the snippet.

Sources: Google: title links (Google Search Central), Google: snippets (Google Search Central), Bing Webmaster Guidelines, §13: titles, descriptions and headings (Microsoft Bing), Microsoft: optimizing content for AI search answers (Microsoft Advertising, 2025-10)

Read today by

Google, Bing, and link previews.

Google Search

Direct

Read the check page
Passed Adopted robots.txt crawl policy — /robots.txt has 1 group and 17 rules across all groups, and 1 Sitemap line.
Observation
```
HTTP 200 · text/plain · 575 bytes · User-Agent: * # Allow robots to be able to crawl the OG Image API route and all the subpaths Allow: /api/og/* Allow: /api/docs-og* Allow: /…
```
Crawlers read robots.txt before they fetch a page (RFC 9309). Its Sitemap line tells them where the URL list is.

Sources: RFC 9309 (IETF, 2022-09), Google robots.txt (Google Search Central)

Read today by

Google, Bing, and the AI crawlers that document robots.txt support.

Google Search

Direct

Read the check page
Passed Adopted XML sitemap — Sitemap https://vercel.com/crawled-sitemap.xml parses and lists 8876 URLs, each with <lastmod>.
Observation
```
HTTP 200 · application/xml · 1209 KB · final https://vercel.com/crawled-sitemap.xml · root <urlset>
```
A sitemap lists the URLs a site wants crawled. Google says a small, well-linked site may not need one.

Sources: Google: build a sitemap (Google Search Central), sitemaps.org (sitemaps.org), Bing Webmaster Guidelines, §3: lastmod (Microsoft Bing)

Read today by

Google and Bing.

Google Search

Direct

Read the check page
Passed Adopted Documentation reachable — Documentation is reachable at /docs (homepage anchor) and returns HTTP 200 text/html with 1368 readable words.
Observation
```
tried: /docs 200 | /docs/frameworks/full-stack/nuxt 200 | /docs/frameworks/full-stack/sveltekit 200 | /documentation 200 | /developers 200 | /gelistiriciler 200
```
Documentation linked as service-doc (RFC 8631), or from the homepage, is how a client learns the API.

Sources: RFC 8631 (IETF)

Read today by

Developers and agents integrating the API.

Google Search

None

Read the check page
Passed Adopted Canonical URL — Canonical https://vercel.com/ matches the page reached.
Observation
```
<link rel="canonical" href="https://vercel.com">
```
A canonical link names the one URL a page is known by. Google treats it and redirects as canonical signals.

Sources: Google: consolidate duplicate URLs (Google Search Central), RFC 6596 (IETF), Bing Webmaster Guidelines, §6: duplicate URLs and canonicals (Microsoft Bing)

Read today by

Google and Bing.

Google Search

Direct

Read the check page
Passed Adopted Structured pricing information — Homepage markup carries 1 complete price node; the first, Offer (unnamed), states 0 USD.
Observation
```
{"price":0,"priceCurrency":"USD","where":"homepage"}
```
A price in markup is read without parsing prose. An Offer with a price and no currency is ambiguous.

Sources: schema.org/Offer (Schema.org), Google: product snippet (Google Search Central)

Read today by

Clients that read schema.org Offer. Google reads an Offer only inside a Product, which product-schema measures.

Google Search

None

Read the check page
Passed Adopted HTTP caching of the homepage — The homepage sends Cache-Control "public, max-age=0, must-revalidate" and ETag, and a request with If-None-Match was answered HTTP 304 Not Modified.
Observation
```
Cache-Control: public, max-age=0, must-revalidate · ETag: W/"mj6uko8rpactyo" · Last-Modified: (none) · conditional GET with If-None-Match: HTTP 304
```
Cache-Control says how long the homepage may be reused. A validator lets a crawler ask whether it changed, and a 304 answer skips the body.

Sources: RFC 9111: HTTP Caching (IETF, 2022-06), RFC 9110, 13: conditional requests (IETF), Google: HTTP caching for crawlers (Google Search Central, 2024-12), Bing Webmaster Guidelines, §3: ETag and lastmod (Microsoft Bing)

Read today by

Googlebot sends If-None-Match and If-Modified-Since. Every HTTP cache reads Cache-Control.

Google Search

Direct

Read the check page
Passed Adopted Language declaration and multilingual access — html lang="en" is a well-formed BCP 47 tag, with no hreflang alternates.
Observation
```
html lang="en" · alternates: none · content-language: (none)
```
The lang attribute names the language of the text. WCAG 3.1.1 requires it at level A. Google reads hreflang, not lang.

Sources: WCAG 2.2, 3.1.1 (W3C), Google: managing multi-regional sites (Google Search Central), Google: localized versions (Google Search Central)

Read today by

Screen readers and translation tools. Google ignores the lang attribute and reads hreflang.

Google Search

Indirect

Read the check page
Passed Adopted Service/product scope machine-readable — Homepage markup declares 1 Service-type node; Service "Vercel" offers "Agentic infrastructure platform" from Vercel Inc. to Worldwide.

Service markup states what is offered, by whom and where. An incomplete node leaves a client to guess.

Sources: schema.org/Service (Schema.org)

Read today by

Clients that read schema.org Service. Google has no Service rich result.

Google Search

None

Read the check page
Passed Adopted Per-page content separation (no SPA fallback) — The homepage and /about have different titles ("Agentic Infrastructure - Vercel" and "About Us") and different main text (8% word overlap).
Observation
```
compared / with /about · titles "Agentic Infrastructure - Vercel" / "About Us" · skeleton 88% · main-text word overlap 8% · main text 1346 characters
```
A route that answers every URL with the same page makes every URL look alike to a crawler.

Sources: Google: HTTP errors and soft 404s (Google Search Central)

Read today by

Google flags such pages as soft 404s.

Google Search

Direct

Read the check page

Hand this report to your coding agent

The prompt lists 8 checks to fix. Each has the value the scan measured and a command that repeats the measurement. It adds the change to make where the scan could build one. After them it lists 4 agent conventions, at lower priority. The prompt is served at https://agenticwebscore.com/aws/vercel.com/prompt.md. The command below prints it, so you can paste the output into your agent.

curl -s https://agenticwebscore.com/aws/vercel.com/prompt.md

report.json

This scan cannot see which AI answers cite the site. Bing Webmaster Tools reports citations in Copilot and in Bing's AI summaries under AI Performance. Google Search Console includes AI Overviews and AI Mode in its Performance report totals, without a separate view (Google's guide to AI features). Both need a verified owner of the site.

Badge

Show this score on your own site. Press Get the badge, paste its snippet into your homepage, and press Check my badge. The directory then lists your site as a verified listing.

Read how listing works

Ship a fix, then re-scan

Measured 8 days ago. Nothing re-scans a site on its own. A scan runs once, when somebody asks for it — and that somebody need not be you. Anyone can request a scan of any public host, so this page may be the first you hear of it. After you ship a fix, press Re-scan now. A check that now passes adds its points to the score.

Why a re-scan can differ

The current rubric evaluates 3 of its 60 checks against a clock: security-txt, status-page, and inventory-freshness. They can change with no change to the site. A response slower than 12 s is reported as not measured, so the same bytes served slowly can move the score and the fingerprint.

Measured as: SaaS / software product

Can an agent understand what the product does, what it costs, and how to call it? Checks that only apply to other site profiles were not scored, so you are not asked to do work that does not apply to you.

· SaaS / software product (score 8), matched: SoftwareApplication schema; A pricing page responds; A documentation path responds; Service schema.

Signal scores

SaaS 8
API platform 4
Services 4
Marketplace 3
Publisher 2

Modules

Every site: always in scope.
Sites with an API or agent surface : in scope.
- · Link rel="api-catalog" </.well-known/api-catalog> on the homepage
- · /.well-known/api-catalog lists 2 linkset entries
- · /.well-known/oauth-authorization-server names issuer https://vercel.com
Sites with a commercial signal : out of scope. The homepage shows no commercial signal.

Scan record

Target: https://vercel.com
Measured at: https://vercel.com/
Scanned: 2026-10-20 17:05:33Z
Requests: 72
Duration: 7.5s
Confidence: 100% · high
Rubric: rubric v0.9

Fingerprint

d50502a02fc9… verify this record

full hash

d50502a02fc9bbfd53dc5ae986a13999b96bc4829f1766a3314c1166080a5ce5

Page as an agent receives it

Words: 537
Est. tokens: 725
Markup: 584.7 KB
Language: en

“Agentic Infrastructure - Vercel”

By dimension

Each score is the share of the in-scope web-standards points a dimension earned. A dash means the dimension has no web-standards check in scope. The agent conventions are scored apart, as their own part.

Discoverability
96

Pass

5

Partial

0

Fails

1

Not counted

0

Unscored: 0 in use, 0 not used
Content
97.3

Pass

8

Partial

2

Fails

0

Not counted

0

Unscored: 0 in use, 1 not used
Crawler access
100

Pass

1

Partial

0

Fails

0

Not counted

0

Unscored: 1 in use, 1 not used
API and agent
47.5

Pass

2

Partial

2

Fails

1

Not counted

1

Unscored: 2 in use, 0 not used, 2 with an error
Commerce
—

Pass

0

Partial

0

Fails

0

Not counted

5

Unscored: 0 in use, 0 not used

Scores by dimension
Dimension	Score	Pass	Partial	Fails	Not counted	Unscored
Discoverability	96	5	0	1	0	0 in use, 0 not used
Content	97.3	8	2	0	0	0 in use, 1 not used
Crawler access	100	1	0	0	0	1 in use, 1 not used
API and agent	47.5	2	2	1	1	2 in use, 0 not used, 2 with an error
Commerce	—	0	0	0	5	0 in use, 0 not used

Request transcript 72 requests

Every fetch this scan record logged, in order. A fetch that followed redirects shows how many redirect requests it made.

56 fetches, 16 redirect requests.

/

200 text/html 584.7 KB 295ms
/robots.txt

200 text/plain 575 B 121ms
/sitemap.xml

200 +1 redirect application/xml 1209.0 KB 159ms
/llms.txt

200 text/plain 5.1 KB 151ms
/.well-known/security.txt

200 text/plain 374 B 120ms
/security.txt

200 text/html 2566.1 KB 1129ms
/about

200 text/html 541.8 KB 32ms
/contact

200 text/html 429.6 KB 108ms
/pricing

200 text/html 1222.1 KB 53ms
/docs

200 text/html 871.5 KB 40ms
/developers

200 +2 redirects text/html 324.7 KB 696ms
/.well-known/api-catalog

200 application/linkset+json 405 B 37ms
/openapi.json

200 application/json 3072.0 KB 95ms
/.well-known/openapi.json

404 text/plain 9 B 139ms
/.well-known/mcp/server-card.json

404 text/plain 9 B 179ms
/.well-known/ai-catalog.json

200 application/ai-catalog+json 2.0 KB 129ms
/.well-known/agent-card.json

404 text/plain 9 B 264ms
/.well-known/oauth-protected-resource

404 text/plain 9 B 247ms
/.well-known/oauth-authorization-server

200 application/json 2.2 KB 43ms
/.well-known/openid-configuration

200 application/json 2.2 KB 53ms
/products/observability

200 text/html 754.5 KB 115ms
/products/previews

200 text/html 595.4 KB 50ms
/llms-full.txt

200 text/html 2478.6 KB 1789ms
/mcp

200 +2 redirects text/html 324.7 KB 999ms
/openapi.yaml

200 text/html 2478.6 KB 1666ms
/api/openapi.json

404 application/json 84 B 302ms
/swagger.json

200 text/html 2478.6 KB 1545ms
/.well-known/agent-skills/index.json

200 application/json 5.7 KB 63ms
/.well-known/agent-skills.json

404 text/plain 9 B 336ms
/.well-known/skills.json

404 text/plain 9 B 304ms
/auth.md

200 text/html 2522.3 KB 1371ms
/.well-known/ard.json

404 text/plain 9 B 438ms
/.well-known/agent-resources.json

404 text/plain 9 B 285ms
/ard.json

200 text/html 2522.2 KB 1291ms
/documentation

200 +1 redirect text/html 871.5 KB 150ms
/gelistiriciler

200 +2 redirects text/html 324.7 KB 797ms
/status

200 +2 redirects text/html 324.7 KB 979ms
/changelog

200 text/html 578.8 KB 115ms
/releases

200 +2 redirects text/html 324.7 KB 957ms
/whats-new

200 +2 redirects text/html 324.7 KB 518ms
/docs/sandbox

200 text/html 861.8 KB 363ms
/sandbox

200 text/html 898.6 KB 83ms
/developers/sandbox

200 text/html 2557.8 KB 1557ms
/contact/sales

200 text/html 504.2 KB 494ms
/get-started.md

200 text/markdown 7.5 KB 103ms
/docs/getting-started-with-vercel.md

200 text/markdown 8.9 KB 611ms
/docs/agent-resources/vercel-plugin.md

200 text/markdown 15.5 KB 354ms
/docs/agent-resources/skills.md

200 text/markdown 13.5 KB 423ms
/docs/products.md

200 text/markdown 2.2 KB 349ms
/docs/sitemap.md

200 text/markdown 490.1 KB 1188ms
/docs/frameworks/full-stack/nuxt

200 text/html 1442.7 KB 262ms
/docs/frameworks/full-stack/sveltekit

200 text/html 1433.3 KB 238ms
/

200 text/markdown 1.7 KB 376ms
/

304 — 0 B 66ms
/aws-probe-3f511295eb8c

200 +2 redirects text/html 16.0 KB 147ms
/aws-probe-3f511295eb8c

404 text/markdown 271 B 69ms

Your HTTP cache result

Verdict

Defects 8

Optional choices 3

Agent conventions 9

Out of scope for this site 24

Sites with an API or agent surface

Sites with a commercial signal

Every site

Passed 16

Hand this report to your coding agent

Badge

Ship a fix, then re-scan

Measured as: SaaS / software product

Scan record

Fingerprint

Page as an agent receives it

By dimension

Discoverability

Content

Crawler access

API and agent

Commerce