AEOCheck.ai
BETA
Live audit · runs against your real site, not a cached snapshot

Audit your site forthe agent web.

ChatGPT, Claude, Perplexity, and Gemini decide whether to cite you. AEOCheck grades your site across the five pillars that matter — including llms.txt and AVL, the new standards for the machine-readable web.

Most audit tools don't even know what AVL is. We do — we ship it ourselves at aeocheck.ai/.agent.

20+AI crawlers checked
llms.txt+ AVL detection
5 weighted pillars · grade A–F
Free · report by email

text/agent-view·v1

We don't just audit AVL.
We ship it.

Agent View Layer is the emerging standard for parallel machine-readable views of every page on your site. Same data, different render — structured @meta, @intent, @state, @actions — so AI agents don't have to reverse-engineer your site from pixels.

Most audit tools don't even know AVL exists. AEOCheck does. We probe five AVL signals on every audit (manifest, agent route, content negotiation, link tag, discovery badge) and grade you against the AVL reference implementation.

GET/.agent↗GET/agent.txt↗AVL spec ↗
aeocheck.ai/.agenttext/agent-view
@meta
version: 1
route: /
ttl: 3600

@intent
purpose: Free website audit for AEO/GEO
audience: site owners, SEO pros, AI agents
capabilities:
  - score across 5 weighted pillars
  - detect llms.txt, AVL, schema.org
  - probe 20+ AI crawlers

@actions
  - method: POST
    path: /api/leads
    purpose: Start a free audit

@nav
self: /.agent
manifest: /agent.txt

Five pillars. One weighted grade.

Every check rolls up into a 0–100 score and an A–F letter grade. The report tells you exactly which checks failed, what to change, and the point cost of each fix.

25% of total score
AI Discovery

What we look at

  • robots.txt + sitemap reachability
  • Allow/Disallow for GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User
  • 20+ AI crawler user-agents in total
If GPTBot or ClaudeBot is disallowed, AI assistants literally cannot read your site to cite it.
25% of total score
Content Structure

What we look at

  • Title tag + meta description quality
  • H1/H2 hierarchy and semantic HTML
  • Answer-capsule pattern in the first paragraph
AI summarizers latch onto the first paragraph. A buried thesis doesn't get cited.
20% of total score
Structured Data

What we look at

  • JSON-LD, Microdata, RDFa parsing
  • Schema.org types: Organization, LocalBusiness, Article, FAQPage, …
  • NAP consistency (name / address / phone)
Schema is how AI knows the difference between a brand mention and a brand entity.
15% of total score
User Experience

What we look at

  • Lighthouse Core Web Vitals (LCP, CLS, INP)
  • Contact info + form completeness
  • Mobile responsiveness signals
Slow pages get downranked everywhere — including in AI overview citations.
15% of total score
AI Communication — llms.txt + AVL

What we look at

  • /llms.txt + /llms-full.txt detection and validity
  • AVL: /agent.txt, /.agent, link rel='alternate', content negotiation, discovery badge
  • Citation likelihood: stats, quotes, authority signals, freshness
The new standards. Most audit tools don't even check for AVL. We do — and we ship it ourselves.