Skip to content
siterank.info
Sign in Start free

Website intelligence for the AI-search era

Understand how search engines and AI systems see your website.

SiteRank AI analyses technical SEO, AI readiness, topical authority, crawler access and observed AI citations in one place — and shows the evidence behind every finding.

Checks one page plus robots.txt, the sitemap and llms.txt. No account needed. Nothing is changed on your site. See how it works

What can actually be measured

  • Crawlable robots.txt, status codes, indexability observed
  • Parseable headings, passages, entities, schema observed
  • Mentioned and cited answers to a prompt set you choose sampled
  • Recommended to a person not measurable from outside the product not claimed

Search changed. Most tools did not.

Conventional search visibility still matters and is still well documented. What is new is that people also get answers from systems that retrieve pages, summarise them and sometimes cite a source. Those are different mechanics, and they need different measurements — not the same rank tracker with a new label.

Crawled An agent is permitted to fetch the page and does.
Retrieved A system selects the page while answering a question.
Used The answer draws on what the page says.
Cited The response attributes the page as a source.
Recommended A reader acts on it. Nobody can measure this for you.

The two shaded stages are the ones SiteRank AI can observe directly. Any product that claims the rest is guessing. Read the full explanation

From SEO to GEO: what changed?

Classic SEO still matters. Search engines and AI answer systems both need pages they can discover, crawl, understand, and trust. What changed is the path: visibility may now involve retrieval, synthesis, mentions, and citations, not only a blue-link ranking.

Technical SEO

SEO foundation Observed per page

What changed
Nothing here went away. Titles, descriptions, canonicals, headings, internal links, indexability, page quality, and performance still affect whether a page can be discovered and understood.
What SiteRank AI looks at
Titles, meta descriptions, canonicals, heading structure, indexability, duplicate or thin content, broken internal links, insecure links, image alt text, and page-level technical issues.
What to improve
Fix crawl blockers, clean up metadata, consolidate duplicate URLs, improve internal linking, and make each page clearly answer a real user need.

Crawler access

Access control Read from robots.txt

What changed
Search crawlers, AI-search crawlers, training crawlers, and user-triggered agents may have different purposes and different robots.txt controls. OpenAI, for instance, documents OAI-SearchBot for ChatGPT Search, GPTBot for model training, and ChatGPT-User for user-triggered fetches as separate agents.
What SiteRank AI looks at
robots.txt, sitemap availability, llms.txt presence, blocked paths, crawler-specific rules, and whether important pages can be fetched.
What to improve
Decide which agents should access your public content, which should be blocked, and whether search visibility and model-training preferences should be handled separately.

Structured data and entity clarity

GEO readiness Inspected, not scored

What changed
AI and search systems need clear facts about the page, organization, author, product, service, article, and topic. Ambiguous pages are harder to retrieve and cite accurately.
What SiteRank AI looks at
Schema.org markup, organization details, author and publisher signals, visible page facts, structured data consistency, and conflicts between plugins or templates.
What to improve
Use accurate structured data, make the brand and topic unmistakable, show authorship and editorial context where relevant, and keep visible content aligned with markup.

Passage-level content

Retrieval-ready Structure observed

What changed
AI systems often retrieve and synthesize passages, not just whole pages. A page may rank, but still fail to provide a clean answer-sized passage.
What SiteRank AI looks at
Heading hierarchy, section clarity, concise answer blocks, topical coverage, internal links, page freshness, and whether content can stand alone without hidden context.
What to improve
Break complex pages into clear sections, answer specific questions directly, support claims with evidence, and connect related pages into useful topical clusters.

Topical authority

Site architecture Computed locally

What changed
Visibility is increasingly tied to whether a site demonstrates depth around a topic, not only whether one page matches one keyword.
What SiteRank AI looks at
Topic clusters, pillar candidates, orphan pages, internal linking gaps, coverage depth, related content, and whether important pages reinforce each other.
What to improve
Build connected topic hubs, update stale pages, remove or improve thin content, and make the site's expertise obvious through structure.

AI mentions and citations

LLM visibility Sampled under recorded conditions

What changed
AI visibility can mean being mentioned in an answer, cited as a source, compared with competitors, or omitted entirely. Mentions and citations are observable, but they are not the same as a traditional rank.
What SiteRank AI looks at
Approved prompt sets, provider used, date observed, answer text, brand mentions, cited URLs, competitor citations, share of voice, and volatility over time.
What to improve
Create source-worthy content with clear facts, original value, useful examples, transparent sourcing, and pages that directly satisfy the questions people ask.

Human trust and action

Trust layer Not measurable from outside

What changed
Even when a page is crawled, retrieved, mentioned, or cited, the final decision still belongs to a person. Trust signals and usefulness still determine whether visibility turns into action.
What SiteRank AI looks at
Transparency signals, freshness, content clarity, evidence, author or organization information, and whether recommendations are tied to observable page issues.
What to improve
Make the page accurate, current, useful, easy to verify, easy to navigate, and clear about who is behind it.

Start with what can be observed: crawl access, technical SEO, structured data, topical coverage, and citation monitoring under recorded conditions.

Run a website check See how SiteRank AI works

One view of website visibility

Six analyses over one content inventory, so a finding in one shows up as context in the others.

Technical SEO

Titles, descriptions, canonicals, heading structure, thin and duplicate content, broken and insecure internal links, orphan pages, image alt text. Every finding names the URL and the value observed.

AI readiness

Entity clarity, machine-readable structure, passage retrievability, authorship and site transparency, freshness. Each check carries its evidence tier, and experimental ideas never affect a status.

AI visibility

Mention rate, citation rate, share of voice and volatility across a prompt set you approve, each reported with its numerator, denominator and confidence interval.

Topical authority

Clusters computed locally from your own content, with pillar candidates, coverage and connectivity bands, orphan pages and internal linking opportunities.

Crawler access

Which search, AI-search, AI-training and user-triggered agents your robots.txt allows, what each operator documents about blocking, and a proposal you approve before anything changes.

Structured data

Schema types found across the site, which plugin produced them, and where two sources describe the same page differently.

Evidence, not AI mythology

Most of what is written about optimising for AI answers is untested. So every recommendation here carries the strength of the evidence behind it, and the weakest two tiers are never counted against you.

Evidence tiers and whether they affect a status
Tier Means Affects a status
ESTABLISHED STANDARD A published specification or settled web standard Yes
OFFICIAL PROVIDER GUIDANCE Stated in current first-party documentation Yes
STRONG EVIDENCE Reproducible, publicly documented measurement Yes
EMERGING PRACTICE Widely practised, plausible, unconfirmed Yes, at reduced weight
EXPERIMENTAL Little or conflicting evidence No
HYPOTHESIS First-principles reasoning only No

Read the methodology

Guides

Written against primary sources, dated, and reviewed when the documentation changes.

Hello world!

Welcome to WordPress. This is your first post. Edit or delete it, then start writing!

All guides

Start with your own site

The SEO, GEO and topic analysis runs entirely on your own installation and needs no API key. Add a provider key later, only if you want visibility monitoring.

Create a free account See how it works