Methodology

How we audit your site.

Every score on Audora.One traces back to a specific check, a specific severity, and a specific point deduction. No black box, no vibes, no "AI said so." This page tells you exactly what we run, how we weigh it, and where the line between the free and paid report lives.

What we test

We run 26 individual checks against your URL, grouped into seven categories. Each check is a discrete piece of code — read it, fork it, file an issue with it.

  • Performance — Lighthouse audits via our self-hosted Chromium runner, plus PageSpeed Insights cross-check, cache + compression headers.
  • SEO — meta tags, canonical, sitemap discovery (4 candidate paths), robots.txt parsing, hreflang, structured data, internal-link crawl.
  • Accessibility — heading hierarchy (skipped levels, missing h1, multiple h1s), alt-text coverage, Lighthouse a11y audits.
  • Security & Trust — TLS handshake (cert validity, hostname match, signature algorithm), HSTS, CSP, X-Content-Type-Options, mixed content scan, subresource integrity, outdated CMS detection.
  • Email & Domain — SPF, DMARC, DKIM, CAA records pulled from your domain's DNS.
  • Social & Sharing — Open Graph, Twitter Card, OG image presence and validity.
  • Conversion — Claude AI semantic pass on your hero copy, CTA strength, information architecture, and how an AI agent (GPT, Claude, Perplexity) would describe your site to its users. Includes llms.txt spec validation and AI-crawler robots policy.

How we score

Every category starts at 100 points. Each finding deducts from that category's score based on its severity:

  • High — minus 18 points. Things that materially break the page or expose users to risk: missing HSTS on HTTPS, mixed content, expired TLS cert, no <h1>, broken viewport meta.
  • Medium — minus 8 points. Real problems with workarounds: missing CSP, missing meta description, multiple h1s, missing sitemap.
  • Low — minus 3 points. Nice-to-fix polish: missing canonical, missing favicon, length-out-of-range title tags, missing AI-crawler rules in robots.txt.

Your overall score is the average of the five user-facing category scores: Performance, SEO, Accessibility, Security, and Conversion. Email and Social are reported separately because they fail in their own ways and most operators want to look at them in isolation.

Findings are also tagged with an impact score from 1–100. This is what drives the order findings appear in the report — and which four make it into the free tier.

Free tier vs. paid tier

The free report shows the top 4 findings ranked by impact score. That's enough to know whether your site has serious problems and roughly where they sit. It is not the full picture.

The paid report ($19, one-time) unlocks:

  • Every finding from every check, ranked and grouped by category
  • The Claude semantic pass — copy clarity, CTA strength, hero-section diagnosis written in plain English
  • The AI Perception check — what GPT-4, Claude, and Perplexity actually say about your site when asked
  • Quick wins, biggest impact, and fast fixes — three pre-sorted action lists so you don't have to triage
  • Multi-page Lighthouse runs (homepage + up to 4 internal pages we discover)
  • A token-protected sharable URL you can send to a developer or stakeholder

What we don't do

We don't run JavaScript-heavy crawls of your full site (yet). We don't audit logged-in pages. We don't tell you whether your content is "good" — we tell you whether it's findable, parseable, and trustable, which is what most audits actually need.

We don't make up numbers. Every threshold on this page is constant in the source: app/Audit/Methodology.php for the severity weights and per-check rationale, app/Providers/AppServiceProvider.php for the registered check list. If a number on your report surprises you, it should be traceable.

Want the per-check detail?

The audit report itself includes a "Methodology" tab on every finding — explaining what we tested, what triggers each severity, and what the impact score curve looks like for that specific check. That's where the depth lives. This page is the map; the report is the territory.