Transparency

How we calculate your AI Search Readiness Score.

Every score is deterministic. Same URL, same HTML, same output. The audit fetches your homepage, /llms.txt, and /robots.txt from our server (User-Agent: TheAeoGeoAuditor/1.0), then applies the signals below. Total possible: 100 points, capped at 100.

Signal
Weight
What we measure
Authoritative source

llms.txt present

+20

HTTP 200 on GET /llms.txt with non-empty body

llmstxt.org spec ↗

robots.txt AI-bot rules

+10

robots.txt mentions GPTBot / ClaudeBot / PerplexityBot / Google-Extended / CCBot / anthropic-ai

OpenAI GPTBot docs ↗

JSON-LD block detected

+8

One or more script type application/ld+json blocks parse as valid JSON

Schema.org JSON-LD ↗

Organization schema

+6

JSON-LD graph contains an @type matching /organization/i

schema.org/Organization ↗

Product / FAQ / Breadcrumb schema

+6

Any of Product, Offer, FAQPage, BreadcrumbList present

Google structured data ↗

Schema diversity (>=3 types)

+5

Distinct @type count >= 3 across all JSON-LD blocks

schema.org type list ↗

Title tag length (20-65)

+5

title length between 20 and 65 characters

Google SEO starter ↗

Meta description length (60-165)

+5

meta name description length between 60 and 165

Google SEO starter ↗

Canonical URL

+3

link rel canonical href present

Google canonicalization ↗

og:image

+2

meta property og:image content present

OpenGraph protocol ↗

Single h1

+6

Exactly one h1 tag on the page

MDN HTML sectioning ↗

>=2 h2 tags

+4

h2 count >= 2

MDN HTML sectioning ↗

Word count >= 300

+6

Whitespace-tokenized text extracted from stripped HTML

internal - RAG chunk floor

Word count >= 800

+4

Same tokenizer as above

internal - long-form bonus

Has main or article

+5

Regex match for landmark tags in raw HTML

HTML Living Standard ↗

Semantic density >= 0.35

+5

chars(main)/chars(bodyStripped)

internal - signal/noise ratio

What we deliberately don't measure yet.

JavaScript-rendered content — we fetch static HTML only, so SPAs with empty body shells score poorly by design.
Interior page depth — v1 measures the homepage only. Multi-URL crawls are on the roadmap.
Actual citation counts in ChatGPT / Perplexity / Gemini — no public API exposes these; anyone claiming to measure them is estimating.
Rendering performance (LCP, INP) — Core Web Vitals are already covered by PageSpeed Insights.
Backlink authority — SEMrush / Ahrefs handle this better.

Reproducibility

Every audit is verifiable in your terminal.

The Evidence Panel on every result shows the exact HTTP status, response time, headers, and raw HTML/JSON-LD snippets we parsed — plus copy-paste curl commands so you can reproduce the fetch with the same User-Agent we used. If our numbers don't match yours, that's a bug and we want to hear about it.