CrawlCheck
AEO audit · free, no signup

AEO audit: can ChatGPT, Claude, Perplexity and Google reach, read and quote your site?

In one sentence

An AEO audit measures, in order, whether an answer engine's crawler can reach a page at all, whether it receives readable words rather than a shell or a challenge, and whether the page holds a specific, consistent fact the engine can repeat with attribution — graded A to F with the fix list in priority order.

Type a domain. Answer-engine optimisation is three questions in order: can the crawler get the page at all, does it get real words rather than a shell or a challenge, and is there a specific, consistent fact it can repeat with attribution. The scan measures all three and grades them A to F.

No signup. Nothing installed. Nothing changed on your site. The same scan as the homepage - about 30 read-only requests - and the same report, opened at the sections that answer this question.

How often an answer engine gets less than a browser
The page is mostly codePAGE_IS_MOSTLY_CODE26.2%An answer engine is refusedANSWER_ENGINE_REFUSED5.4%Every crawler refused alikeUNIFORM_REFUSAL10%A challenge page served as 200CHALLENGE_SERVED_2000.5%The words need JavaScriptCONTENT_NEEDS_JAVASCRIPT1.4%A crawler served less than a browserCRAWLER_SERVED_LESS0.5%Share of 4,310 scans in the public dataset that carried the finding. Live; the same figures as /data.

Is your site one of them? Check it free → All 63 rates →

Measured

What the scan measures for this

Reach

GPTBot, ClaudeBot, PerplexityBot, Google-Extended and the rest: what each was served at the edge, and what robots.txt told it.

See this section on a real report: Crawler permissions →

Read

Visible text as a share of the page, the machine files that parse, and the sitemap that resolves - the page as a crawler receives it, not as a browser renders it.

See this section on a real report: Page weight & text →

Quote

A lead sentence that defines the subject, paragraphs quotable whole, question headings that are answered, FAQ text that is actually visible, sentences carrying a checkable fact.

See this section on a real report: Quotable content →

Consistency

Name, phone, address, hours and the claims the page makes, checked across the page, its JSON-LD, its entity map and the directories that list it.

See this section on a real report: Claim consistency →

For agents

What an agent is told about the same site

An agent does not read the report. It asks one endpoint before it reads, cites, connects or transacts, and gets a signed answer built from the same measurement.

1 · See it

The agent view

The answer an agent receives about any domain, rendered for a person: each action’s decision and the reason.

Agent view

2 · Call it

One endpoint

GET /api/v1/resolve?domain= returns the signed answer; preflight returns a decision and a receipt for one action.

QuickstartResolve spec

3 · Check it

Without trusting us

Every answer is Ed25519-signed and logged; it verifies offline with the published key.

Signing keysProof

Workflow

The workflow

  1. Scan the domainAbout thirty read-only requests, the fifteen identities sent in the same second; about twenty seconds for a site new to us. No account, nothing installed.
  2. Read the three stage ringsReach, Read and Quote each carry a score; the weakest names what holds the grade down. The Fix list is already in priority order.
  3. Fix in that orderReach defects first - a refused crawler makes every other row moot. Then readability. Then the copy: the Quotable section prints the bar each row is held to.
  4. Re-scan, then compareThe second scan is the proof. Compare against a competitor to see which stage separates you.
Free

Free tools for this

Each one is a single request to the domain you name, answered as text or a download. Put the domain on the end of the URL, or use the form on the tools page.

Reading

Read next

Start from what you are checking

You cannot pay to improve a score. A licence adds the lists behind the counts and the record over time - never a better grade.

Questions about this page

QWhat is AEO?
Answer-engine optimisation: making a site the source an engine such as ChatGPT, Claude, Perplexity or Google's AI answers can reach, read and quote. The scan measures those three stages directly rather than inferring them from rankings.
QHow is this different from an SEO audit?
An SEO audit asks how a page ranks; this asks what a crawler received. A page can rank and still ship to GPTBot as a 403 or a script shell, and the answer engine never sees it. Every row here is a measurement of the delivered bytes.
QWhat does the Quote stage actually score?
Whether the first sentence defines the subject, the share of paragraphs quotable whole, whether question headings are answered, whether FAQ schema text is actually visible on the page, sentence length, and the share of sentences carrying a checkable fact — each row prints the corpus figure the bar came from.
QCan I pay for a better AEO grade?
No. Nothing on the pricing page changes a measurement. A licence adds the lists the free report holds back and the dated record; the one-time fix service changes the site, and the change is re-measured rather than asserted.