Audit details, section by section

Audit details is the third tab of the Site Report. Nothing here is a recommendation. It is the evidence the scores and fixes were built from, so you can see what Serpio observed before you act.

Five sections, one badge each

Every section is collapsed by default. Click the header to expand it. The badge uses the same four levels as the score cards: Strong, Good, Needs work, Critical, plus Not enough data when there is no signal. A View recommendations button at the top jumps back to the fix list.

Site health

This is the only section built entirely from deterministic checks. The Site health snapshot card shows a score out of 100. Next to it, Confidence reads High, Medium, or Low, with a line saying how many pages were sampled and how they were found: sitemap, internal links, or a homepage-only fallback. Fewer than three pages triggers a shallow-sample warning.

What this result means gives a one-line summary and a reminder that the score summarises crawlability, page structure, entity clarity, citation cues, and agent access. It is not a traffic, ranking, or uptime score. Expand How the score is calculated to see the five components. Crawlability and Structure each count for a quarter of the total, Citation readiness for a fifth, and Entity clarity and Agent accessibility for the rest.

Crawlability

Sitemap found, canonical coverage across the sample, no duplicate titles or descriptions, hreflang and llms.txt as small bonuses.

Structure

Title, meta description, a single H1, several H2s, structured data, Open Graph tags, and a viewport tag on the homepage.

Entity clarity

Whether the homepage says who you are: a descriptive title, a substantive description, an H1, and Organization or WebSite schema.

Citation readiness

Schema present, FAQ or HowTo markup, canonical tags, and enough paragraph text for a passage to be quoted.

Agent accessibility

Whether a non-browser client can read the page: metadata, canonical, schema, real paragraph text, and an llms.txt file.

Evidence to inspect lists the findings with P1, P2, or P3 tags. The first three show by default and Show more findings reveals up to seven. Each finding expands into Why it matters, Fix, and Owner. What machines can identify rates five things High, Medium, or Low: Organization identity, Offering clarity, Answer-ready content, Machine-readable structure, and Citation cues. Analysis coverage gaps lists anything the analysis could not check.

Search engines and AI citation readiness

These two sections come from the written analysis of the sampled pages, so treat their notes as observations rather than measurements. Search engines (SEO) shows Overall Health and Internal Link Profile, then On-Page Fundamentals with a signal and a note for Title Tags, H1 Structure, and Meta Descriptions. Keyword Strategy lists Core Targets and Thematic Clusters the analysis inferred from your pages. An Actionable Recommendations box lists the area's suggestions and missing signals.

AI citation readiness shows Overall Readiness and Citation Likelihood, then Structured Data Quality, Entity Clarity, and a Knowledge Panel Signals row labelled System Detection Match. All of these are qualitative labels about how legible your pages are to a machine.

Local & regional

This section only matters if you serve specific places or languages. It shows a Regional Signal and a Local SEO Signal, then three checks under Regional Architecture: Multi-Region Target (Active or Inactive), Hreflang Tags (Detected or Missing), and LocalBusiness Schema (Present or Missing). Coverage & Focus lists the Targeted Regions the analysis inferred, or None detected.

hreflang

A link tag that tells search engines which language or regional version of a page to show. Only add it when you actually have those versions.

LocalBusiness

A schema.org type for a business with a physical location or service area. A global SaaS with no premises does not need it, and Missing is not a problem in that case.

Tech basics

The header row shows four badges. Performance comes from a mobile Lighthouse run requested through Google's PageSpeed Insights API and is rated Strong, Moderate, or Weak from the performance category. Security Grade is a letter from A to D based on how many of five headers the homepage returned: Content-Security-Policy, Strict-Transport-Security, X-Frame-Options, X-Content-Type-Options, and Referrer-Policy. Four or more is an A, three a B, two a C, and anything less a D. Accessibility is Strong, Moderate, or Low from a count of images without alt text, form controls without labels, and pages without a lang attribute. It is not a WCAG audit.

Ecosystem & Stack lists Detected Technologies, a Stack Breakdown (Frontend, CMS, Commerce, CDN / Hosting), and Notable Integrations. Detection uses Wappalyzer when configured; otherwise it reads the generator tag, script URLs, and response headers, and a coverage gap says so. Core Checkups shows the vitals LCP, CLS, FCP, and TBT, yes-or-no rows for lazy loading, image optimisation, caching headers, and the security headers, and Mobile Readiness, where Appears Responsive is inferred from a Viewport Meta Tag.

If a row says Unknown, the check could not run. Unknown never lowers a badge. For the fixes, go to Working through the Recommendations tab, and to When to re-analyze your site once they are deployed.