What the audit grades
Forty-five checks across eight dimensions: Foundational, Technical, On-page, AI Overviews (AIO), Generative Engine Optimization (GEO), Local, E-E-A-T, and Performance. Each check produces a finding with the exact evidence (HTTP headers, JSON-LD blocks, content excerpts, URLs) and a fix shaped as an engineer-ready ticket. We crawl up to 50 pages on the free tier.
What's different about this audit
Most free SEO audits stop at the technical layer. They tell you your meta description is too long and your CLS is 0.18. Real findings, but they don't tell you whether ChatGPT can read your page, whether Google's AI Overview will cite you, or whether Perplexity's passage-extraction pipeline finds your answers. We grade all three explicitly.
The eight dimensions
- Foundational: titles, meta descriptions, canonicals, robots.txt, sitemap, indexability.
- Technical: HTTPS, redirects, Core Web Vitals (LCP, INP, CLS) [3], hreflang, broken links.
- On-page: heading hierarchy, semantic HTML, content depth, intent match, internal linking.
- AI Overviews: answer-first openings, question-shaped H2s, factual density, TL;DR placement above the 800-pixel fold.
- GEO: entity clarity, Wikidata sameAs edges, schema depth (Article, Person, Organization, FAQ).
- Local: NAP consistency, LocalBusiness schema, geo coordinates, locations pages.
- E-E-A-T: Person schema for authors, credentials, external citations, About/Contact/Privacy presence.
- Performance: actual page weight, render-blocking assets, modern image formats, lazy loading.
Why "free" matters here
The freemium-with-no-signup format exists because the audit's job is to give you the answer, not to capture your email. Seobility built a $5M+ ARR business on exactly this shape [1], and they were right to. The only honest version of a free SEO audit is one that runs without a signup gate, produces a shareable URL, and tells you everything wrong with your site even if you never come back. Ours does the same, with the AI-search layer added.
What a 2026 audit must include that a 2020 audit didn't
AI crawler access parity
Most sites silently block OAI-SearchBot, PerplexityBot or ClaudeBot at the robots.txt or WAF layer without realising it [4] [5]. If GPTBot can read you but ChatGPT's search index (OAI-SearchBot) cannot, you are invisible to ChatGPT search. We test each one.
Passage-level extractability
AI engines retrieve passages, not pages. The optimal extractable unit is roughly a 134-167 word self-contained answer block immediately after a question-format heading. 44.2% of all LLM citations come from the first 30% of a page. We grade whether your pages are shaped that way.
Entity density and schema depth
Pages with 15+ Knowledge Graph entities show 4.8x higher selection probability in AI answers. Real people, real places, real organizations, with Person and Organization schema linking via sameAs to Wikidata, LinkedIn, GitHub. We test all of it.
What you get
- Score 0-100 across all eight dimensions.
- Executive brief (non-technical) and technical brief (engineer-ready).
- 30/60/90 plan with owners, effort and impact.
- Raw evidence for every finding.
- Shareable public URL (or keep it private).
- PDF export available on paid tiers.
How to read your score
Above 80 is strong; 60-80 means optimization potential; below 60 is critical. The score is a guide, not a verdict - the findings list is the operational truth. Start at the top of the prioritized list and ship one finding at a time. Most sites pick up 10-15 points in the first sprint.