Scoring methodology
The NiceCite Score is an evidence-based grade of how prepared a website is to be quoted by AI answer engines. It is calculated from a live crawl of your pages and six signal categories — not from third-party rank data, not from estimated metrics, and not from anything we cannot point to in your HTML.
What it measures: Whether pages directly answer the questions a real user (or an LLM) would ask, near the top, in language an answer engine can lift verbatim.
Pulled from your site: H1 specificity, presence of direct-answer paragraphs, FAQ blocks, list/table structure, and quotable sentence density across the audited pages.
What it measures: How clearly the site names what it is, who it is from, and what it does — so an LLM can attribute a citation without ambiguity.
Pulled from your site: Explicit brand/product names in titles and H1s, About-page presence, contact/identity exposure, consistent naming across pages, and Organization signals.
What it measures: Machine-readable identity and content markup that helps AI engines map your pages into a knowledge graph.
Pulled from your site: JSON-LD blocks parsed per page (Organization, WebSite, SoftwareApplication, FAQPage, BreadcrumbList, Article, Product, LocalBusiness), flattened across @graph and array @type values.
What it measures: Verifiable, third-party-checkable signals that an AI engine can use to decide whether to cite a page.
Pulled from your site: External links to authoritative sources, presence of author/contact info, social profile linkage, HTTPS, and consistency between visible content and structured data.
What it measures: Whether the site has enough crawlable, on-topic content for an AI engine to form a confident answer.
Pulled from your site: Word counts, heading depth, internal link density, presence of topic-defining pages (features, pricing, FAQ, methodology), and per-page content uniqueness.
What it measures: Whether AI crawlers and classic search bots can actually reach and parse the pages.
Pulled from your site: robots.txt status, sitemap.xml presence and validity, canonical tags, HTTP status of audited URLs, and the share of audited pages that returned clean HTML.
Every audit begins at the URL you provide. NiceCite fetches robots.txt, reads sitemap.xml when present, and follows internal links from the homepage. Pages are crawled up to your plan's page limit. The complete URL list, the per-page HTTP status, and the parsed signals are exposed in the Score Integrity panel of every report so you can verify exactly what produced the score.
NiceCite respects robots.txt, does not log in, and does not bypass authentication. Pages behind auth, paywalls, or noindex are skipped.
The NiceCite Score is deterministic — same site, same crawl, same score. It changes when something underneath it changes: a page you edited, schema added or broken, a canonical tag fixed, new URLs added to your sitemap, or a plan change that allows more pages per audit. Each report stores the scoring version it was generated with so historical comparisons stay honest across formula updates.
Free accounts get a real audit, all six category scores, the prioritized fix plan, and a shareable report. Premium increases pages per audit, unlocks historical tracking and competitor comparison, adds monitoring, and includes generation credits for FAQ drafts, schema, and on-page rewrites. The scoring formula is identical on every plan.
Yes. Given the same crawl snapshot (same URLs, same HTML, same schema), the score is deterministic. Scores change when your live site changes or when the audited page set changes — for example, a plan upgrade that crawls more URLs.
No. No tool can guarantee that ChatGPT, Gemini, Perplexity, Claude, or Google AI Overviews will cite a specific page. NiceCite measures readiness signals those engines are known to rely on and gives you an evidence-based fix plan.
All scores and evidence are measured from your live HTML: crawled URLs, schema blocks, headings, body text, meta, sitemap, robots. Generated outputs (FAQ drafts, schema snippets, on-page rewrites) are AI-assisted suggestions clearly labeled as such, and they never feed back into the score.
NiceCite starts from your homepage, reads sitemap.xml and robots.txt, follows internal links, and crawls up to your plan's page limit. The exact URL list is shown in the report's Score Integrity panel so you can verify what was scored.
Either your site changed (content added/removed, schema fixed or broken, new canonicals) or the audited URL set changed (plan limit, new sitemap entries). The Score Integrity panel diffs the page list and shows the scoring version used.
Because the formula, category weights, and signal definitions are ours. The current scoring version is published in each report so you can see exactly which version produced a given number.
Every signal the NiceCite audit grades against, mapped to the exact route, section, or structured-data block on this site that satisfies it, and the verification method a third party can use to confirm it. This is documentation, not a scoring hack — nothing here changes any score.
| Signal | Where on NiceCite | Verify via |
|---|---|---|
| Clear, specific H1 with audience + outcome | Homepage hero | Raw HTML |
| FAQ block with question-style headings | Homepage FAQ section + /faq | Raw HTML |
| FAQPage JSON-LD | Homepage @graph | JSON-LD |
| Direct-answer paragraphs above the fold | Homepage answer blocks | Raw HTML |
| Signal | Where on NiceCite | Verify via |
|---|---|---|
| About page exists | /about | Public link |
| Organization JSON-LD with stable @id | Sitewide @graph | JSON-LD |
| Business name reinforced on homepage | Hero, FAQ, footer | Raw HTML |
| Signal | Where on NiceCite | Verify via |
|---|---|---|
| Organization, WebSite, WebPage | Sitewide @graph | JSON-LD |
| SoftwareApplication + Offer | Homepage @graph | JSON-LD |
| Service schema for the audit | Homepage @graph | JSON-LD |
| FAQPage schema mirrors visible FAQ | Homepage + /faq | JSON-LD |
| BreadcrumbList per page | Per-route head() | JSON-LD |
| Signal | Where on NiceCite | Verify via |
|---|---|---|
| Public methodology page | /methodology | Public link |
| Public trust / evidence page | /trust | Public link |
| Contact page | /contact | Public link |
| Editorial / methodology team byline | /methodology, /trust | Raw HTML |
| Crawl evidence panel inside every report | Report viewer | Crawl evidence |
| Signal | Where on NiceCite | Verify via |
|---|---|---|
| Service / product / pricing pages | /features, /pricing, /aeo-checker, /ai-citation-audit | Public link |
| Topical depth on AEO and methodology | /answer-engine-optimization, /methodology | Public link |
| FAQ + About + Contact coverage | /faq, /about, /contact | Public link |
| Signal | Where on NiceCite | Verify via |
|---|---|---|
| sitemap.xml lists all public pages | /sitemap.xml | sitemap.xml |
| robots.txt allows crawlers | /robots.txt | robots.txt |
| Canonical + meta description on every public page | Per-route head() | Raw HTML |
| Server-rendered HTML (not SPA-only) | TanStack Start SSR | Raw HTML |
Crawlable H1, FAQ schema, public methodology, trust pages, answer-ready copy, and evidence-based reports — the same checks NiceCite runs on every audited site.
Maintained and reviewed by the NiceCite methodology team. Last updated May 2026.
Run a free NiceCite audit and see every signal scored against the six categories above, with the evidence behind each number.