200 audit checks + 9 recurring monitors
Find what holds your site back. Fix it. Prove it worked.
KinetixSEO connects technical SEO, AI citation readiness, rankings and page-specific fixes in one repeatable workflow—so you always know what to do next.
No account · Real findings · Start with any public page
See 200 checks on your site
Get your SEO and AI-readiness baseline
Enter a domain or page. We’ll connect the signals and show what deserves attention first.
200 checksSEO, GEO and diagnostics
9 monitors5 nightly + 4 weekly
6 AI engineslive citation tracking
Verified fixesre-check the original finding
The product loop
From uncertainty to measurable improvement
Features matter only when they move the work forward. KinetixSEO turns them into a clear four-step loop.
01 · Find
See what blocks visibility
Audit technical SEO, content, schema, crawlability and AI citation readiness together.
02 · Prioritise
Know what matters first
Rank findings by severity, impact and effort instead of receiving another flat checklist.
03 · Fix
Apply tailored changes
Get grounded replacement copy or copy-paste code based on the page that was checked.
04 · Verify
Prove the issue cleared
Re-run the original checks and see verified, not applied, or regressed.
What you can achieve
One platform for visibility problems that overlap
Audit and improve · 200 checks
Make every important page easier to find, understand and trust
Combine classic SEO and infrastructure diagnostics with passage-level AI readiness in one prioritised report.
- ✓149 classic SEO checks
- ✓44 AI-readiness checks
- ✓7 infrastructure diagnostics
- ✓Page-specific fixes and direct verification
Monitor and defend · 9 routines
Know when Google or AI engines stop choosing you
9 recurring routines keep watch after the audit: 5 run nightly and 4 run weekly.
- ✓Google rankings, opportunities and competitor gaps
- ✓6 AI engines checked for live citations
- ✓Independent citation ground-truth cross-checks
- ✓Brand accuracy, mentions and change alerts
Create and expand
Build pages grounded in facts—not generic filler
Turn keyword opportunities into brand-consistent landing pages with claims grounded against your own site.
- ✓Research-led landing page generation
- ✓Batch and programmatic pages
- ✓Fact ledger and brand voice grounding
- ✓Published-page verification
Diagnose precisely
Use a focused tool when you already know the question
Run one piece of the audit engine for schema, sitemaps, speed, security, AI agents and more.
- ✓24 focused tools
- ✓No signup or email
- ✓Same underlying audit logic
- ✓A fast answer for one known issue
Monitoring in full
9 recurring routines keep watch after the audit
5 run nightly and 4 run weekly. The cadence and descriptions come directly from the customer product catalogue.
Weekly AI visibility check (included)
Once a week, each keyword you track for AI visibility is checked on every AI engine we measure, using its first prompt. The current engine list is on the features page; it is read from what is actually switched on, never assumed. Included in your plan: this does not spend credits.
AI visibility prompt checks (extra prompts and on-demand runs)
A keyword's first prompt is checked every week for free. Each extra prompt you add (up to 3 per keyword) is checked on every AI engine in the same weekly run and uses credits each week. The "Run AI visibility check" button runs every prompt right away and uses credits for each one, including the first. If your balance runs out, extra prompts are skipped until you top up — your first prompt keeps running.
AI brand accuracy check
Once a week we ask an AI model what it knows about your brand from its own training, with no web search, and record whether it confuses you with someone else and how accurate its description of you is. This is what an AI answers about you when nobody is looking it up. Included in your plan — this does not spend credits.
Brand mention tracking
Once a week we search the web for pages that mention your brand without linking to you, and check each one to confirm whether a link is really missing. You get a ready-made outreach list of sites that already know who you are. Included in your plan — this does not spend credits — and you can switch it off per site.
Rank tracking
Every keyword with rank tracking switched on gets a fresh Google rank check, every single night, at zero credits and zero clicks. A dedicated rank-tracker is one more tool and one more bill elsewhere — here it's just part of the plan. What you pay for is how many keywords you can track, never for the checks themselves.
Keyword opportunity refresh
Search volume and difficulty refresh for every rank-tracked keyword, every night, on their own, so the opportunity scores on your Rank Tracking tab are never stale. No manual refresh, no credits spent finding out what changed, included in every plan.
Keyword gap refresh
Your keyword gap matrix against tracked competitors rebuilds every night on its own — the moment a competitor starts ranking for something you don't, it's already there next time you look. No manual "refresh" button exists for this on purpose (see admin notes): a rebuild is already running for you, every night, at no credit cost.
Search Console sync (automatic)
Search Console data (impressions, clicks, average position) and Bing data pull in on their own every night for connected sites on a paid plan. Each pull covers a rolling 28-day window ending three days ago, because Google keeps revising the most recent days. Free: Google doesn't charge for this API, and neither do we.
AI referral import (automatic)
Once you connect Plausible or Google Analytics, every night we import how many visits reached your site from ChatGPT, Perplexity, Claude and other AI assistants, next to your search engine visits. Search Console never shows this traffic. We read daily visit counts per source and landing page only, never individual visitors. Free for connected sites on a paid plan.
AI readiness and AI visibility remain separate: readiness checks whether an engine can cite you; monitoring records whether it actually does.
Why KinetixSEO
Not another dashboard that leaves the work to you
The difference is the closed loop between evidence, action and verification.
“A score tells you where you are. A verified fix tells you whether you moved.”
Every paid Page Fix includes one direct verification re-scan against the original findings.
Claims stay bounded by evidence
Generated fixes and pages are grounded against facts found on your site. Signal confidence is shown instead of hiding uncertainty.
Useful standards, honest limits
KinetixSEO can generate llms.txt for assistants that fetch it, while stating clearly that Google Search ignores it. Housekeeping is not sold as a ranking shortcut.
Choose your starting point
Start free. Pay when you want the fix or ongoing proof.
I have one page or one question
Run the complete free check or choose a focused tool. No account required.
Check a page free →Best for ongoing growth
I want fixes, monitoring and history
Choose a plan for Page Fix credits, recurring monitoring, site-wide workflows and saved progress.
Compare plans →For technical evaluators
Every check behind the 200-check audit
Explore the complete live catalogue. Counts and descriptions are generated from the same registries used by the scanner.
Classic search visibility
SEO · 149 checks
Technical, on-page and content signals that determine whether search engines can crawl, understand and rank a page.
Content QualityWhether a page says enough, says it clearly, and says something a reader could not get from a thinner competitor — thin content, readability, duplication, and filler language.10
- criticalContent couldn't be analyzedThe scanner couldn't parse the page's content at all, which usually means the URL isn't returning real, crawlable HTML.
- criticalThin contentPages with too few words rarely cover a topic in enough depth to rank well or fully answer what a visitor came looking for.
- highLow readabilityLong, complex sentences and heavy jargon make content harder to scan and understand, for both readers and ranking systems.
- mediumAI-generic filler languageGeneric filler phrases like "let's dive in" or "unlock the power of" read as low-effort AI writing and dilute concrete, useful content.
- highDuplicate content riskNear-duplicate pages compete against each other for the same rankings and dilute the signals search engines use to pick a winner.
- mediumPaywalled or gated contentContent locked behind a subscribe or sign-in wall can go unindexed unless flexible-sampling structured data tells search engines what is behind it.
- mediumH1 doesn't match page contentA heading that doesn't reflect what the body actually discusses is a topical-relevance red flag for both readers and search engines.
- lowLow text-to-HTML ratioA page that's mostly markup with little visible text gives search engines very little substantive content to evaluate.
- mediumNo specific claims or dataConcrete numbers, dates, and sourced facts read as more trustworthy and citable than vague, generic statements.
- lowNo first-hand experience signalsLanguage showing real testing or direct experience ("I tested", "we found") signals the content reflects genuine use, not just aggregated research.
Technical SEOThe foundation search engines need before content quality even matters: crawlability, indexability, HTTPS, redirects, robots directives, and security headers.40
- criticalPage returns an HTTP errorA page returning a 4xx or 5xx status cannot be indexed at all — it needs to load successfully or redirect properly to the right destination.
- highUnresolved redirectInternal links and canonical references pointing through a redirect hop instead of the final URL waste crawl budget and dilute link equity.
- criticalNot served over HTTPSHTTPS is a confirmed Google ranking signal, and browsers actively warn visitors away from HTTP pages, hurting trust and conversions.
- criticalNoindex directive presentA noindex tag or header removes the page from search results entirely, even if everything else about it is fine.
- criticalrobots.txt blocks all crawlersA site-wide Disallow rule stops search engines from crawling any page at all, silently taking the whole site out of search.
- criticalPage blocked by robots.txtA page-specific Disallow rule can block Googlebot from an important page just as completely as a site-wide block, and is often shipped by accident.
- highrobots.txt returns a server errorWhen robots.txt itself fails, Google halts crawling of the whole site for hours and falls back to a stale cached copy for up to 30 days.
- mediumSlow server response timeA slow time-to-first-byte delays everything downstream — page rendering, Core Web Vitals, and how much of the site crawlers can get through.
- highSoft 404 pageA missing page that returns HTTP 200 instead of a real 404 wastes crawl budget and confuses search engines about what actually exists.
- mediumText compression not enabledServing HTML, CSS and JS uncompressed sends several times more bytes than necessary on every single request, which slows Largest Contentful Paint for every visitor and is one of the cheapest fixes available — a single web-server directive.
- mediumInconsistent www/non-www redirectWhen both domain variants don't consistently redirect to one canonical version, search engines can split ranking signals between the two.
- lowInconsistent trailing-slash handlingServing both a trailing-slash and non-trailing-slash URL as separate live pages creates avoidable duplicate-content confusion.
- highMissing security headersWithout headers like CSP, HSTS, and X-Frame-Options, the site is more exposed to clickjacking, downgrade attacks, and injected content.
- mediumMixed content on HTTPS pageResources loaded over plain HTTP on an HTTPS page trigger browser warnings and can be silently blocked, breaking the page for visitors.
- highForm submits over insecure HTTPSubmitting form data — including credentials or personal details — to a plain http:// endpoint exposes it to interception in transit.
- mediumCookies missing security flagsWithout Secure, HttpOnly, and SameSite, a cookie can be sent over plain HTTP, read by injected JavaScript, or attached to cross-site requests — each one a step toward session hijacking or CSRF.
- lowServer version disclosed in response headersA response header naming the exact server/framework version hands an attacker a shortlist of known vulnerabilities to try against this site.
- criticalSite flagged as malicious by Google Web RiskA site on Google's threat list gets "Dangerous site" browser warnings that block most visitors outright and can trigger a search ranking penalty — this is the single most damaging finding a scan can surface.
- mediumMissing XML sitemapWithout a sitemap.xml, search engines have to rely purely on link discovery to find and prioritize the pages worth crawling.
- mediumMalformed sitemapA sitemap that isn't well-formed XML is typically ignored entirely by search engines, providing no crawl-discovery benefit at all.
- mediumRequires JavaScript to renderSome crawlers and link-preview bots don't execute JavaScript, so they see a near-empty page unless it's rendered server-side.
- criticalLeaked credential in page sourceA live API key or credential visible in HTML or JS is publicly exposed to anyone who views source and should be rotated immediately.
- mediumHreflang tag issueInconsistent or missing hreflang declarations can send search engines the wrong language or region variant for a given searcher.
- lowInvalid image sitemap entryAn <image:image> block without a required <image:loc> gives search engines no URL to actually fetch the image from, so the entry is silently ignored.
- lowInvalid video sitemap entryA <video:video> block missing required fields like thumbnail, title, description, or a content/player location gives search engines too little to index the video, so the entry is silently ignored.
- highUncrawlable linksLinks built with javascript: hrefs, #-only anchors, or empty hrefs can't be followed by crawlers, hiding whatever they point to.
- highRisky redirect signalCross-domain redirects and instant meta-refreshes read as manipulative signals rather than a legitimate same-site 301 redirect.
- mediumCross-domain canonical URLA canonical pointing at a different domain tells search engines the authoritative version of this content lives elsewhere entirely.
- lowNo analytics tool detectedWithout an analytics tool installed, there's no way to measure traffic, conversions, or search performance for this page.
- mediumUnsafe target="_blank" linksA target="_blank" link without rel="noopener" lets the destination page access window.opener and redirect the original tab (reverse tabnabbing).
- lowExposed mailto email addressPlain-text mailto links are easy targets for automated spam-harvesting bots scraping email addresses off the page.
- highDead-end page with no linksA page with zero internal or external links gives users nowhere to go next and crawlers no way to discover more of the site from it.
- mediumLinks point to localhostA leftover dev or staging link to localhost is broken for every real visitor and crawler outside the developer's own machine.
- criticalOutbound link to a malicious domainLinking out to a domain on Google's Web Risk threat list can itself trigger browser security warnings and damage visitor and search-engine trust, even though the flagged content lives on someone else's site.
- mediumContent built with APIs Google can't runGoogle's indexing renderer doesn't support IndexedDB, requestIdleCallback or permission-gated device APIs, so anything your scripts only draw after one of those resolves is never seen or indexed.
- lowThird-party scripts without integrity checksScripts loaded from someone else's server with no integrity hash run whatever that server sends, so a compromised CDN can inject anything it likes into your pages.
- lowClickable elements that aren't real buttonsA div with a click handler is invisible to keyboard users, screen readers and the AI agents that increasingly operate sites for visitors, because all of them look for the button role rather than the handler.
- mediumInvisible but clickable elementsAn element nobody can see but everybody can click is a usability trap, and search engines treat invisible-yet-active content as a deception signal.
- lowJavaScript-dependency checks didn't runThese checks weren't collected on this scan, so nothing is being claimed about them either way.
- lowAgent-interaction checks didn't runThese checks weren't collected on this scan, so nothing is being claimed about them either way.
On-Page SEOThe classic ranking signals on the page itself — titles, meta descriptions, headings, canonical tags, and Open Graph data.22
- criticalMeta tags couldn't be analyzedThe page couldn't be parsed for meta tags at all, which usually means the URL isn't returning real, crawlable HTML.
- criticalMissing title tagThe title tag is one of the strongest on-page ranking and click-through signals — without one, the page has no clear SERP headline.
- mediumTitle tag too shortA title that renders far narrower than the 580px Google allows wastes SERP real estate, and usually fails to carry enough context or the primary keyword to attract clicks.
- lowTitle tag too longGoogle truncates the desktop title past about 580px of rendered width — not at a fixed character count — so a title of wide characters can be cut off well before 60 characters, losing its most compelling part.
- highMissing meta descriptionWithout a meta description, search engines generate their own snippet from page text, which is often less compelling than a written one.
- lowMeta description too longGoogle truncates the desktop snippet past about 920px of rendered width — not at a fixed character count — cutting off part of the summary before it can persuade a click.
- mediumMissing canonical tagWithout a canonical URL, search engines have to guess which version of a page is authoritative, risking duplicate-content dilution.
- lowGeneric anchor textLink text like "click here" or "read more" tells users and search engines nothing about the destination, unlike descriptive anchor text.
- lowOutdated meta keywords tagGoogle has ignored the meta keywords tag for ranking since 2009, and its presence reads as an outdated or spam-adjacent SEO approach.
- criticalMissing H1 headingThe H1 is the clearest on-page signal of a page's main topic to both search engines and readers scanning the page.
- mediumMultiple H1 headingsMore than one H1 muddies the single-topic hierarchy a page should present, making the primary subject less clear.
- lowDeprecated HTML tagTags like <center> or <font> have been obsolete in HTML for years and should be replaced with modern CSS equivalents.
- highHeading structure couldn't be checkedWithout parseable HTML, H1 presence and heading hierarchy can't be verified at all, hiding a potentially serious structural issue.
- lowMissing Twitter Card tagsWithout twitter:card meta tags, links to this page won't preview correctly when shared on X/Twitter, hurting social click-through.
- lowNo social profile linksOfficial social profile links feed the Organization schema's sameAs property, which helps Google disambiguate this entity in its Knowledge Graph and confirm it's the same organization elsewhere on the web.
- mediumNon-descriptive URL pathA URL built from numeric IDs or query strings gives users and search engines no readable clue about the page's content before clicking.
- mediumPage type mismatch with SERPWhen a page's format doesn't match what's currently ranking for its target keyword, it's competing against a content type Google prefers there.
- mediumMissing html lang attributeWithout a lang attribute, accessibility tools, translation prompts, and language-targeted search results can't reliably identify the page's language.
- mediumDuplicate canonical tagsMultiple conflicting canonical tags confuse search engines about which URL is actually authoritative for this content.
- lowMissing og:title tagWithout an og:title tag, the page has no clear title when shared on social media, hurting how links appear when posted.
- lowMissing og:description tagWithout an og:description, shared links fall back to arbitrary page text instead of a meaningful, chosen description.
- mediumMissing og:image tagLinks shared without an Open Graph image get far less engagement than ones with a representative preview image.
Structured DataWhether your JSON-LD schema markup is present, valid, and current — deprecated or malformed schema can silently lose rich results.13
- lowStructured data couldn't be checkedWithout parseable HTML, structured data markup can't be verified at all, hiding any schema issues that might exist.
- highNo structured data foundWithout JSON-LD markup, the page misses out on rich results and gives search engines no explicit signal about what type of content it is.
- highDeprecated schema typeA no-longer-supported schema type provides no ranking or rich-result benefit and should be removed or replaced.
- mediumSERP-retired schema typeGoogle no longer generates a rich result for this schema type, so it can't be relied on for search-result appearance anymore — but the markup stays valid schema.org and still helps AI systems understand and cite the page.
- mediumSchema uses HTTP @contextA JSON-LD block declaring an http:// @context instead of https:// is an easy, low-effort schema-hygiene fix.
- mediumMissing required schema fieldA schema block missing a required field is incomplete: for types Google still rewards, it disqualifies the page from that rich result entirely; for every type, it weakens how reliably AI systems can parse and cite the page.
- mediumInvalid breadcrumb markupBreadcrumbList markup needs sequential positions, names, and item URLs to actually qualify for the breadcrumb rich result.
- mediumLow-quality FAQ markupGoogle no longer shows FAQ rich results (removed May 2026), but AI answer engines still read FAQPage markup as question-answer pairs — so every question needs a real name and a substantive, non-duplicate answer to be worth citing.
- lowClient-side rendering blind spotIf schema markup is only injected via JavaScript, it may never reach the served HTML that non-JS-executing crawlers actually see.
- mediumArticle headline too longGoogle disqualifies an Article schema block from its rich result entirely once the headline exceeds roughly 110 characters — it isn't wrapped, just dropped.
- mediumArticle schema missing imageGoogle requires an image of at least 1200x800px for Article-family rich results, so a missing image blocks eligibility outright.
- mediumArticle schema image too smallAn Article schema image below Google's minimum dimensions fails the rich-result eligibility bar even though an image is present.
- mediumArticle schema violationThe Article schema block fails one of Google's structural requirements, blocking it from rich-result eligibility until adjusted.
Core Web VitalsReal-user loading and interactivity metrics — LCP, CLS, FCP, TTFB, and INP — the speed signals that affect both ranking and conversion.6
- lowCore Web Vitals not measuredUntil a real Core Web Vitals measurement runs, there is no data on how the page actually performs for real users on LCP, CLS, or INP.
- criticalPoor Largest Contentful PaintA slow LCP means visitors wait too long to see the page's main content load, which is both a UX and a Google ranking signal.
- criticalPoor Cumulative Layout ShiftContent that jumps around as the page loads frustrates users and can cause accidental clicks on the wrong element.
- mediumPoor First Contentful PaintA slow first paint leaves visitors staring at a blank screen longer than necessary before anything renders at all.
- mediumPoor Time to First ByteA slow server response delays every subsequent rendering milestone, dragging down the whole page-load experience.
- mediumPoor Interaction to Next PaintSlow response to clicks and taps makes a page feel sluggish and unresponsive even after it has visually finished loading.
AI Search ReadinessWhether a page is structured so AI assistants can find, parse, and safely cite it — frontloaded answers, entity density, freshness signals, and llms.txt.14
- mediumNo front-loaded answerAI engines prefer a direct answer in the first 40-60 words of a section — content that builds up to its point is less likely to be lifted and cited.
- mediumLow entity densityNamed entities like people, places, organizations, and brands are what AI systems use to ground and cite content — generic terms give them nothing to anchor to.
- lowNo statistics detectedNumeric evidence — specific dates, counts, percentages, or measured results — improves how likely AI systems are to cite a passage as reliable.
- mediumNo freshness signalWithout a visible published/updated date and matching structured data, AI search systems have no recency signal to score the content on.
- lowNo author signalAI systems weigh clear authorship as a citation-worthiness signal, so content with no visible byline is less likely to be trusted and cited.
- mediumAI readiness signals unavailableWithout parseable HTML, AI search readiness signals like entity density and answer-first structure can't be assessed at all.
- mediumMissing llms.txt fileAn llms.txt file is a low-effort, standardized way to guide AI crawlers to a site's most important content per the llmstxt.org spec.
- lowllms.txt missing site nameWithout a valid H1 site name as the first line, AI crawlers reading llms.txt may not reliably identify which site it belongs to.
- highAI crawlers blocked in robots.txtBlocking AI crawlers like GPTBot or ClaudeBot in robots.txt prevents this content from ever being cited by AI assistants at all.
- lowSite files unavailableWithout a successful domain-level fetch, llms.txt and robots.txt AI-crawler access can't be checked for this scan.
- mediumBelow passage-level citability analysisPresence checks (does an answer-first paragraph exist? are there enough named entities?) are a weak proxy for real AI citability — the deeper passage-level analysis found this content less citable than the checklist alone suggests.
- lowNo Google "Preferred Sources" buttonGoogle's Preferred Sources button lets readers mark a recurring-publisher page as a preferred source for Top Stories/AI Overviews/AI Mode — a new (Aug 2026) feature with no measured AI-citation impact data yet, so treat it as low-effort and low-priority, not a proven ranking lever.
- mediumContent is going staleAI assistants and search engines both favour recently reviewed pages, and studies of citation behaviour consistently find refreshed pages cited more often than ones left to age.
- lowContent age wasn't measuredThis check wasn't collected on this scan, so nothing is being claimed about how old the page is.
ImagesAlt text, image presence, and the accessibility and SEO signal a page loses when images are missing or unlabelled.7
- lowImages couldn't be checkedWithout parseable HTML, image alt text, formats, and loading attributes can't be audited at all.
- lowNo images on pageRelevant visuals — screenshots, diagrams, product photos — can improve engagement and give AI and image search another way to reference the page.
- highMissing image alt textAlt text helps screen readers describe images to visually impaired visitors and can drive additional traffic through image search.
- mediumLegacy image formatOlder formats like JPEG and PNG are typically 25-50% larger than WebP or AVIF at equivalent quality, slowing down page load.
- criticalLazy-loaded hero imageLazy-loading the above-the-fold hero image delays exactly the element Largest Contentful Paint measures, directly hurting that Core Web Vital.
- mediumMissing responsive image srcsetWithout a srcset attribute, browsers always load the largest image version regardless of viewport size, wasting bandwidth and slowing load.
- mediumVideo issue detectedProblems with embedded video, such as missing captions or a broken embed, can limit both accessibility and citability of that content.
Authority & Trust SignalsE-E-A-T signal strength and internal linking health — orphan pages and thin internal linking quietly undermine authority even when the content itself is fine.14
- lowTrust signals unavailableWithout parseable HTML, spam and trust checks like cloaking, hidden text, and E-E-A-T signals can't be evaluated at all.
- mediumOrphan page riskA page with no internal links pointing to it, and that isn't the homepage, is hard for both crawlers and users to discover through normal navigation.
- lowFew internal linksA page with only a handful of internal links pointing to it has weaker internal-linking signals than better-connected pages on the site.
- criticalFails YMYL E-E-A-T standardsFor Your-Money-or-Your-Life topics like health or finance, content without a named expert author or that reads as unreviewed AI output can be effectively disqualified from ranking.
- highWeak E-E-A-T signalsWeak trustworthiness, expertise, authoritativeness, or experience signals in the content make it a harder sell for search engines to rank highly, especially on sensitive topics.
- criticalPrompt injection aimed at AI readersText that tells an AI assistant how to describe or recommend the page is treated as manipulation by search engines and AI providers alike, and can get the page dropped outright.
- mediumAI-manipulation patterns in the contentUnattributed authority claims, inflated evidence and stacked superlatives read as an attempt to game an AI summariser rather than inform a reader, and they erode trust with human visitors too.
- mediumTracking loads before consentUnder GDPR and similar laws, loading analytics or advertising scripts before the visitor opts in is the violation — asking afterwards doesn't fix it, and it's the part regulators actually fine.
- infoPossible full-screen overlay on loadWe detected a fixed-position, near-fullscreen overlay with a dismiss control. This is only a heuristic — verify it isn't blocking primary content for most visitors before acting on it. If it is shown to mobile visitors immediately on arrival from search, it can hurt both rankings and bounce rate.
- mediumMissing trust pagesA privacy policy, a contact page and an about page are the basic evidence that a real business is behind a site, and both search engines and AI assistants look for them before recommending one.
- mediumLink-farm patternA page with an unusual number of outbound links, nearly all pointing at a single domain, reads as a paid or reciprocal link arrangement, which is a link-scheme violation.
- lowAI-manipulation checks didn't runThese checks weren't collected on this scan, so nothing is being claimed about them either way.
- lowCompliance checks didn't runThese checks weren't collected on this scan, so nothing is being claimed about them either way.
- lowOutbound-link concentration wasn't measuredThis check wasn't collected on this scan, so nothing is being claimed about the page's outbound links.
Spam & Local Business SignalsTwo different risks in one bucket: abuse patterns that can trigger a manual action (cloaking, hidden text, keyword stuffing, link spam, scaled content), and — where relevant — local-business completeness like NAP consistency, hours, and reviews.23
- criticalHigh cloaking riskServing different content to Googlebot than to real users is a manual-action risk that can get a site penalized outright.
- mediumPossible cloaking signalPage behavior that differs by user agent is worth reviewing, since unintentional cloaking still carries the same penalty risk as deliberate cloaking.
- highHidden text detectedHiding keyword-stuffed content from users while showing it to crawlers is a manual-action risk under Google's spam policies.
- mediumKeyword stuffing suspectedRepetitive keyword or location lists read as manipulative to both readers and Google's spam detectors, rather than as genuine, useful content.
- highHigh doorway page riskThin content with high link density and geographic title variation is a classic doorway-page pattern that Google's spam policies specifically target.
- mediumSome doorway page signalsThin content, a canonical pointing elsewhere, or unusual title variation are early signals of a page existing mainly to funnel visitors elsewhere.
- highHigh link spam riskAn excessive external-link ratio or unmarked affiliate links violate Google's link-scheme guidelines and put the page at manual-action risk.
- mediumSome link spam signalsElevated footer or sidebar link density and unmarked affiliate links are worth reviewing before they escalate into a full link-spam risk.
- mediumUGC links missing rel attributesLinks in comments or forums need rel="ugc" or rel="nofollow" so search engines don't attribute user-submitted links' authority to the site.
- highThin affiliate contentGoogle's guidelines require affiliate pages to add real value beyond just linking to merchants — too little original content around multiple affiliate links risks a ranking penalty.
- mediumAffiliate links without review languageAffiliate links without first-hand-experience or review language read as a plain link list rather than a genuine, helpful review.
- highScaled/low-value content riskLow readability, repetitive paragraphs, and no named entities are the fingerprint of mass-produced content that Google's spam policies specifically target.
- mediumSome scaled-content signalsLow readability or a lack of named entities can make content read as mass-produced even before it crosses into full scaled-content risk.
- criticalMalware signals detectedObfuscated scripts are a strong indicator of a site compromise and need immediate investigation before anything else about the page matters.
- mediumSuspicious third-party iframesAn iframe embedding content from an unrecognized or untrusted domain is a common vector for malware and compromised-site injections.
- lowThird-party author bylineAn author byline linking to a different domain than the one being scanned needs confirming as a genuine syndication arrangement, not a reputation-abuse pattern.
- mediumUndisclosed sponsored contentSponsored or partnership language without a corresponding rel="sponsored" link fails to properly disclose paid content to both readers and search engines.
- mediumLow E-E-A-T signal countMissing trust markers like an author byline, About/Contact links, or Organization schema are especially costly on YMYL topics where trust signals carry real ranking weight.
- mediumIncomplete NAP dataIncomplete Name/Address/Phone data in local business schema hurts local-pack ranking eligibility, even when the schema type itself is present.
- lowMissing opening hoursWithout an openingHours property, search engines can't show accurate business hours in local search results.
- lowMissing geo coordinatesLocal business schema without latitude/longitude coordinates reduces eligibility for map and local-pack placements.
- mediumReferences a removed GBP featureMentioning a Google Business Profile feature that has since been removed misleads readers about capabilities that no longer exist.
- lowLow review countReview count and recency both factor into local ranking and review-rich-result eligibility, and this page falls below the informal threshold some platforms use.
AI search readiness
GEO and AEO · 44 checks
Whether AI assistants can reach, parse, trust and quote the page directly.
Brand Authority SignalsWhether an AI system can tell who wrote this and trust the brand behind it — author attribution, entity consistency, named entities, and organization schema.6
- mediumMissing author/organization attributionA visible author or organization byline, mirrored in structured data, is what lets AI systems attribute content to a named, credible source.
- mediumInconsistent brand/entity namingWhen a brand name doesn't appear consistently across title, meta description, and H1, AI systems have a harder time reliably associating the page with that brand.
- mediumNo external brand presence foundA brand that a third-party source like Wikipedia or Wikidata recognizes is a verifiable authority signal AI systems can check independently of anything the page itself claims — on-page content alone can't substitute for it.
- highMissing named entitiesNaming specific people, companies, products, or places is what AI systems use to ground and cite content, rather than generic terms.
- mediumMissing organization schemaOrganization (or Brand/LocalBusiness) structured data gives AI systems and search engines a machine-readable identity to attribute content to.
- mediumMissing statistics or numeric claimsSpecific, verifiable numbers make claims easier for AI systems to trust and cite compared to vague, unsupported statements.
CitabilityThe passage-level traits that make a paragraph quotable by an AI answer — answer-first structure, evidence-backed claims, ideal length, freshness, and low hedging.12
- highNo early, front-loaded answerRewriting key sections to lead with the direct answer, rather than building up to it, makes content far more likely to be lifted and cited by AI systems.
- mediumStale contentContent untouched for 180+ days, and especially a year or more, reads as increasingly stale to AI systems weighing recency in what to cite.
- mediumUnsupported superlative claimsClaim words like "best", "proven", or "guaranteed" without a specific number, stat, or citation nearby are far less likely to be cited by AI answer engines.
- mediumNo freshness signalA visible published/updated date, matched to datePublished/dateModified in structured data, is what AI search systems use for recency scoring.
- mediumAnswer not front-loaded / off sweet spotAI citations concentrate heavily in the first 30% of a page, and total length in the roughly 800-1500 word range has the strongest empirical AI-extraction coverage.
- highPassages not sized for citationPassages in the 134-167 word range are long enough to carry context but short enough for an AI system to quote cleanly as a self-contained answer.
- lowHigh hedge-word densityHedging language like "might", "could", or "it seems" makes claims less likely to be lifted as a confident, quotable statement by an AI system.
- lowNo attributed quotationsQuotations attributed to a named person or organisation give an AI answer engine a statement it can cite with a source, which plain unattributed prose does not.
- mediumStale statistics or outdated techCiting dated studies as current, or referencing long-obsolete technology, undermines the credibility of a passage an AI system might otherwise cite.
- mediumPassages too long for citationLong, sprawling paragraphs are harder for an AI system to extract and quote cleanly than passages broken into a citable, self-contained length.
- mediumLow statistical densityContent dense in concrete, sourced facts and certifications is measurably more likely to be cited by AI answer engines than purely descriptive text.
- mediumStatistics not sourced, dated or on-topicA figure only helps an AI answer when it names its source and year and belongs to the section it sits in; unattributed or off-topic numbers are not evidence an engine can cite.
Multi-Modal ReadinessWhether a page gives an AI system more than plain text to work with — data tables, images, and video content it can also draw on.5
- lowNo data tableA structured data table presents tabular information — pricing, specs, comparisons — in a form that is easy for both readers and AI systems to extract.
- mediumSubstantive text exists only inside imagesSteps, statistics, pricing or explanations that are baked into an image with no text equivalent on the page are invisible to every text-based crawler, including the AI bots behind ChatGPT, Perplexity and Google AI Overviews. That content cannot be cited.
- lowText-in-image check not runImage text detection needs an AI vision pass that is disabled for this scan (free check, or vision disabled by the administrator). The images were not inspected.
- lowToo few imagesRelevant visuals like diagrams, screenshots, or photos give AI and image search another way to reference and cite the page.
- lowNo video contentA relevant video, where it genuinely fits the content, is a strong multi-modal signal that broadens how the page can be discovered and cited.
Structural ReadabilitySemantic HTML that helps an AI system parse a page correctly — proper landmarks, lists, sections, and genuine question-and-answer structure.7
- lowMissing semantic <article> elementWrapping core content in a semantic <article> element gives crawlers and AI systems a clear structural signal about where the primary content lives.
- mediumHeadings styled with div/span instead of h1-h6Text that only looks like a heading (bold, large, coloured) but is not an h1-h6 tag disappears as a heading the moment an AI crawler converts the page to plain text or Markdown. The page's outline collapses into one long paragraph, and answer engines lose the section boundaries they use to pick a passage to quote.
- lowNo lists for scannabilityBulleted or numbered lists break up scannable information like steps, features, or comparisons in a form both readers and AI systems parse easily.
- mediumMissing semantic <main> landmarkA semantic <main> element clearly marks the primary page content, helping both accessibility tools and content-extraction systems locate it.
- mediumQuestion headings not directly answeredQuestion-style headings whose following text doesn't directly and immediately answer the question are less likely to be lifted cleanly by an AI system.
- mediumNo question-per-section structureFraming key topics as a question subheading with the answer directly beneath it matches exactly the shape AI answer engines extract from.
- mediumMissing semantic <section> elements<section> elements dividing distinct topics give crawlers and AI systems a clearer structural map of the content than undifferentiated markup.
AI Crawler AccessWhether AI crawlers and citation bots can actually reach and read a page at all — bot access, llms.txt, markdown negotiation, and content-signal headers.14
- highAI crawlers blocked in robots.txtBlocking AI crawlers in robots.txt directly prevents this content from being citable by whichever AI systems are disallowed.
- highAI-crawler User-Agent blocked or challengedA WAF or CDN (e.g. Cloudflare Bot Fight Mode) can block or interstitial-challenge an AI-crawler User-Agent at the edge even when robots.txt explicitly allows it — a silent, invisible way to lose all AI-citation traffic with nothing showing up in a robots.txt review.
- mediumNo AI-crawler content negotiationRecognizing an AI-crawler user agent and serving it a lighter, stripped-down response is an emerging competitive edge for AI citation, though most sites don't do this yet.
- lowMissing ai.txt fileAn ai.txt file (root or /.well-known/) declares AI training/scraping/indexing permissions — a young, competing-draft convention with no ratified spec and no confirmed AI-crawler adoption, so it's a low-effort bet rather than a proven lever.
- highCitation-intent AI bots blockedBlocking bots that answer live user queries right now, like PerplexityBot or ChatGPT-User, has a high, immediate impact on real-time AI-answer-engine visibility.
- mediumContent-Signal opts out of AI retrievalA robots.txt Content-Signal directive of ai-input=no explicitly opts out of the real-time AI retrieval mechanism that AI-Overview-style answers pull live content through.
- lowNo RSS/Atom feed autodiscoveryA <link rel="alternate" type="application/rss+xml"> tag is the standard way feed readers and AI crawlers discover a site's feed; only relevant for content/blog sites that actually publish one.
- mediumMissing llms.txt filePerplexity, Claude and coding agents fetch llms.txt, but Google Search officially ignores it and log studies across 300k domains found ~97% of published files are never fetched at all — a cheap, low-risk addition per the llmstxt.org spec, not a ranking lever.
- lowNo Accept: text/markdown negotiationHonoring an Accept: text/markdown request header lets AI crawlers request clean content directly instead of parsing it out of full HTML.
- lowNo AI-markdown-twin routeServing a plain-Markdown twin of a page at {path}.md is a cheap, high-signal convention that lets AI agents fetch clean content without HTML stripping.
- highContent only appears after JavaScript runsAI crawlers do not execute JavaScript - OpenAI disclosed that none of roughly 500 million GPTBot fetches rendered the page - so anything assembled client-side is invisible to every AI answer engine, however good it is.
- lowMissing RSL licensing fileA /.well-known/rsl.json file declares content-licensing terms for AI training — an emerging, low-effort signal for sites with a clear licensing stance.
- lowMissing TDMRep fileA /.well-known/tdmrep.json file declares text-and-data-mining reservation status per the W3C TDMRep Community Group Final Report — more standardized than ai.txt, but still has no confirmed adoption by AI crawlers.
- highTraining-intent AI bots blockedBlocking model-training crawlers has no immediate citation impact and may be a deliberate content-rights choice rather than something that needs fixing.
The underlying infrastructure
Diagnostics · 7 checks
Full sub-reports for performance, security, certificates, DNS and page resources.
Infrastructure & DiagnosticsChecks that don't reduce to a single pass/fail — each one runs a full sub-report (certificate details, DNS records, a crawl, a resource-by-resource inventory) rather than a scored finding, but every one of them runs on every full analysis at no extra cost.7
- SSL CertificateIssuer, validity window, key strength, and expiry, checked live — so a lapsed or weak certificate never catches you off guard with a browser warning that tanks trust and conversions.
- Email & DNS TrustSPF, DMARC, MX, BIMI, DKIM, DNSSEC, and domain expiry, checked live against your DNS — the records that decide whether your mail even lands in an inbox.
- Broken Link CheckingEvery link on the page is followed and checked — dead ends and redirect chains flagged before a visitor hits them.
- Static Resource InventoryEvery script, stylesheet, and image weighed by size and origin, with missing cache headers and likely-unminified files flagged.
- Raw vs. Rendered HTMLCatches when client-side JS silently rewrites your title, canonical, robots meta, or H1 after the page loads — invisible in view-source, very visible to a crawler.
- Load Waterfall & ScreenshotsThe order and timing every resource loads in, plus a before/after screenshot pair — useful for spotting late-loading content or render-blocking requests.
- Accessibility (WCAG)A full axe-core scan of the rendered page — every WCAG violation found, categorized critical to minor, down to the exact element and, for color-contrast failures, the measured ratio versus what's required.
See what your page needs next.
Start with a real 200-check baseline. No signup, no sales call, no generic demo.
Check my site free →