Find the site issues that keep you out of answers

Find the problems on your site that stop AI from reading and recommending your pages.

Answers collected from

  • ChatGPT
  • Claude
  • Gemini
  • Google AI Overviews

Site Health highlights

  • Page-level evidence

    Each finding shows the page, what was captured and why it matters for that type of page.

  • AI crawler access

    robots.txt permissions by crawler purpose: search, AI search, training and user-triggered fetches.

  • Coverage beside every score

    Each score shows how much of your site it covers, so you can see when a crawl was partial.

Site Health features

Two scores: Web Fundamentals and AEO Readiness

Web Fundamentals covers the basics every page needs, such as indexability, HTTPS, titles, canonicals, language and the mobile viewport. AEO Readiness scores what answer engines rely on across seven pillars. Only checks that affect visibility or performance count toward a score. The rest appear as advice.

  • Crawlability, machine readability, structure and answerability
  • Evidence, provenance and freshness
  • Each page classified, so only relevant checks apply
  • Advisory and diagnostic checks shown without scoring

Technical and answer-readiness checks together

CiteLadder checks each page for server-rendered content, structured data that is valid and matches what the page says, and clear headings. A check that could not run is marked unresolved, so you know what was not tested.

  • Grouped issues with every affected page
  • Structured data validated against visible content
  • Broken internal links, sitemap and hreflang defects
  • Site-wide duplicate titles, missing H1s, dead-end pages and HTTP that never redirects to HTTPS
  • Internal links: inbound links, depth from home and anchor text

Know which AI crawlers you let in

Site Health reads your robots.txt and reports access for each known crawler by purpose. Blocking a search crawler such as OAI-SearchBot or Claude-SearchBot is scored, because it can keep you out of answers. Blocking training crawlers is your choice, so it is shown without a penalty.

  • Search, AI search, training and user-fetch crawlers kept apart
  • robots.txt history across crawls
  • llms.txt reported as a diagnostic, not scored

From a finding to a verified fix

A technical defect that repeats across pages, usually from a shared template, is one Action covering every affected page, not one per page. Hand it to the Agent with the failing check and the affected pages so it writes one fix, or prepare a developer ticket with the dated evidence. Your team implements it, marks it done, and the next comparable crawl reports it verified, partially resolved, still failing or inconclusive.

  • One Action for a defect across all its pages
  • Developer ticket in Markdown, with a minimal patch
  • Pages still failing named after the recrawl
  • CSV and Markdown exports for your developers

How it works

  1. Crawl

    CiteLadder samples your site within your plan’s page allowance, on demand or on a weekly or monthly schedule, respecting robots.txt unless you choose otherwise.

  2. Classify

    Each page is classified by kind, so a product page and a blog post get the checks that fit.

  3. Review

    Read grouped issues, scores and coverage, and open any page for the captured evidence.

  4. Fix and recrawl

    Brief the fix, ship it, and confirm it in the next crawl.

Questions you can answer

CiteLadder answers each one from the AI answers and site data it has recorded.

  • Is anything blocking our key pages from being indexed or cited?

  • Do our robots.txt rules keep out the AI search crawlers we want in?

  • Does our structured data match what the page actually says?

  • Which page types score worst for answer-readiness, and why?

  • Did last sprint’s fixes actually land?

Frequently asked questions

What does Site Health check?

Technical fundamentals such as indexability, HTTPS, canonicals and titles; answer-readiness across crawlability, machine readability, structure, answerability, evidence, provenance and freshness; structured data; internal links; and robots.txt access for known crawlers.

How is it different from a standard SEO audit?

It adds answer-readiness checks and AI crawler access to the technical basics, classifies each page before checking it, and ties every finding to captured evidence that the Agent can use to write a fix brief.

Does a higher score guarantee AI citations?

No. It summarizes the checks that applied to your pages. Engines decide what to index and cite on their own.

How many pages are crawled?

Up to your plan’s monitored page allowance. Pages are sampled across your site in a reproducible way, and the coverage figure tells you how much was checked.

Can crawls run on a schedule?

Yes, on every plan. Turn on a weekly or monthly schedule for a project and each run repeats the latest crawl’s scope and settings. Schedules are off until you turn them on.

Does the crawl respect robots.txt?

Yes, by default. For your own site you can run a crawl that ignores robots.txt after confirming you are authorized to crawl it; crawl delay, rate limits and access controls still apply, and pages fetched despite an exclusion are marked. Every other fetch, such as competitor, cited-source and product pages, always respects robots.txt.

Is robots permission the same as crawler traffic?

No. Site Health reads your policy. Allowing a crawler is not evidence that it visited. AI crawler logs show the requests that actually arrived.

Does Site Health check llms.txt?

Yes, as a diagnostic. It is reported with your crawl, but it does not affect your score.

Does it measure Core Web Vitals?

No. Site Health reports server response time as a diagnostic, but it does not produce field Core Web Vitals.

Hand your developers a fix list with the evidence attached