Run a website audit

A website audit crawls 5 to 50 pages of your site, scores them on GEO and AEO, and rolls the results into four pillars: Authority, Evidence, Structure, and Freshness. You get a ranked list of fixes with copy-paste code snippets and a PDF you can hand to a developer. It tells you why engines do or don't cite you, and what to change first.

Before you start

You need a website URL. The audit reads your live site, so make sure the pages you care about are public and not behind a login. Scoring adapts to your detected site type: a publisher leans toward GEO, a developer-tools company toward AEO.

Run the audit

Enter your site

Open Website health and paste your URL (for example acme.com). AppearIn detects your site type so it can weight the criteria that matter for how you actually get found.

Let AppearIn crawl

AppearIn crawls 5 to 50 pages, starting from your homepage and following internal links. It reads your content, structure, schema markup, and technical signals. A typical crawl finishes in a couple of minutes.

Read the GEO and AEO scores

Each page is scored on GEO (Generative Engine Optimization, 11 criteria such as content structure, schema markup, citation worthiness, freshness, and E-E-A-T) and AEO (Agentic Engine Optimization, 12 criteria such as documentation structure, quickstart, code examples, llms.txt, and machine-readable sitemaps). See GEO and AEO for what each one means.

Roll up into the four pillars

The individual criteria roll up into four pillars: Authority, Evidence, Structure, and Freshness. The pillars are the fastest way to see where you're weak without reading every line item.

Work the prioritized recommendations

Each finding comes with a fix and, where relevant, a code snippet you can paste in. Start at the top of the list: recommendations are ordered by impact, so the first few are where you'll move the most. Export the PDF for a 30-day action plan you can share with a developer or client.

How to read the pillar scores

Each pillar answers a different question an engine is implicitly asking before it cites you:

Authority: do you look like a credible source? Driven by E-E-A-T signals, author and organization markup, and how citation-worthy your claims are.
Evidence: can an engine verify what you say? Driven by citations, data, examples, and structured facts it can extract cleanly.
Structure: can a model parse you? Driven by headings, schema markup, documentation structure, code examples, llms.txt, and machine-readable sitemaps.
Freshness: is the content current? Driven by visible dates, update cadence, and whether stale pages are pruned.

Prioritize the lowest pillar first, then the highest-impact item inside it. A weak Structure score usually pays back fastest because the fixes are concrete and mechanical, which is where the schema snippets in your recommendations come in.

A worked example: a recommended schema fix

Say the audit flags an article that has no Articlemarkup, hurting both your Structure and Authority pillars. The recommendation includes a JSON-LD snippet like this to drop into the page's <head>:

json-ld

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "How AI engines pick which brands to cite",
  "datePublished": "2026-06-01",
  "dateModified": "2026-06-07",
  "author": {
    "@type": "Person",
    "name": "Jordan Lee"
  },
  "publisher": {
    "@type": "Organization",
    "name": "Acme"
  }
}
</script>

That single block helps three criteria at once: it adds schema markup (Structure), names a named author and publisher (Authority), and exposes a visible dateModified (Freshness). For a help center or FAQ page, the audit recommends a FAQPage block instead, with each question and answer as a structured pair.

What good looks like

A healthy audit result looks like:

No pillar far behind the others. Balanced scores beat one strong pillar masking a weak one.
Schema markup on the page types that earn citations: articles, products, FAQs, and how-tos.
Visible, recent dates on content that should stay current.
An llms.txt and machine-readable sitemap if you ship docs or developer tools.
The top three recommendations resolved before the next crawl, so you can measure the lift.

Common mistakes

Chasing a perfect score on every criterion.Scoring adapts to your site type. A publisher won't need an llms.txt, and a docs site won't live or die on E-E-A-T. Fix what matters for how you get found, not the whole checklist.

Fixing line items instead of pillars. One low criterion rarely moves anything. Look at which pillar is weakest and clear the high-impact items inside it.

Auditing once and walking away. The audit is a baseline. Re-crawl after you ship fixes so the pillar scores reflect the change, and treat the data as a trend.