AgentEye
Docs

How the AgentEye scan works

Everything the free Agent Readiness scan checks, how the score is calculated, and how to use the scan API and badge.

The scan, step by step

  1. 1

    You enter a URL. No signup, no login, nothing installed on your site.

  2. 2

    We fetch up to ~16 public URLs: your homepage, llms.txt, llms-full.txt, .well-known/llms.txt, robots.txt, sitemap.xml, and common pages like /about, /docs, /pricing, /faq, /contact, /terms and /privacy.

  3. 3

    We read the raw responses the way a text-based agent would — no JavaScript execution, no cookies, no deep crawling.

  4. 4

    Every check passes or fails, points add up to a 0–100 Agent Readiness Score, and you get a permanent shareable report at /report/[your-domain].

Score bands: 0–30 Not Agent Ready · 31–60 Partially Agent Ready · 61–80 Mostly Agent Ready · 81–100 Agent Ready.

Scoring model — 7 categories, 100 points

Agent Discovery

20 pts
  • llms.txt exists10
  • llms-full.txt or extended markdown resource exists5
  • .well-known/llms.txt exists3
  • llms.txt contains useful links/headings2

Crawlability & Robots

15 pts
  • robots.txt exists4
  • robots.txt does not block all crawlers5
  • homepage returns status 2003
  • important pages accessible without login3

Sitemap & Structure

10 pts
  • sitemap.xml exists5
  • sitemap contains multiple useful URLs3
  • canonical tags or clear meta structure2

Content Clarity

20 pts
  • homepage has clear title and meta description4
  • homepage contains a clear product/company description5
  • about page exists3
  • FAQ or explanatory content exists3
  • content is not only hidden inside JavaScript5

Docs / API Readiness

15 pts
  • docs page exists4
  • developer or API page exists4
  • OpenAPI/spec/reference links found4
  • markdown or text-based docs available3

For local businesses this category is replaced by Local Business Signals (address, opening hours, phone, menu, booking).

Trust & Canonical Facts

10 pts
  • company/contact/legal pages exist3
  • privacy/terms pages exist3
  • official links/socials are clear2
  • canonical facts are easy to extract2

Risk, Claims & Compliance

10 pts
  • risk/disclaimer language exists3
  • no excessive unsupported claims detected3
  • pricing/limitations are clear2
  • deprecated/outdated info is marked clearly2

Scan API reference

Run a scan programmatically. This OpenAPI-style reference covers the public scan endpoint — no API key required, fair-use limits apply.

POST https://agentready-eosin.vercel.app/api/scan
Content-Type: application/json

{
  "url": "https://example.com"
}

Response

{
  "reportUrl": "/report/example.com"
}

The report page renders the score, category breakdown and full diagnosis. A dynamic share image is available at /api/report-card/[domain].

Badge embed

After a scan, every site gets a live SVG badge that shows its current score and links back to the report:

<a href="https://agentready-eosin.vercel.app/report/your-domain.com">
  <img src="https://agentready-eosin.vercel.app/badge/your-domain.com" alt="Agent Ready score" />
</a>

Machine-readable resources

We practice what we scan. This site publishes its own agent-readable resources in Markdown and plain text:

See how your site scores against all of this

Run a free scanRead the FAQ
Docs — How the AgentEye scan works