Transparent scoring
Audit methodology
Every score is a bounded summary of observable website signals. Findings and evidence—not the number alone—should drive changes.
What the audit measures
Focused tools separately assess crawler access, discoverability, raw-page extractability, structured entities, content clarity, citation evidence, first-party brand identity, and agent readiness. The benchmark averages access, discoverability, extractability, entity clarity, evidence, and freshness.
Rules and severity
Deterministic checks use versioned rule IDs. Critical and high findings identify access, reading, discovery, interpretation, or trust problems. Medium and low findings identify improvements and best practices. Experimental signals do not reduce stable scores.
Score interpretation
Missing data is neutral when absence is not a confirmed defect. Every deduction is represented by a concrete finding.
Scan context and limits
- Scans are point-in-time and usually inspect one submitted page.
- Discoverability samples at most eight sitemap URLs.
- Standard scans read initial HTTP responses and do not run JavaScript.
- Private networks, credentials, non-standard ports, oversized responses, and long redirects are blocked.
- Readiness does not guarantee an AI mention, recommendation, ranking, or citation.
How to reproduce a finding
- Open the affected public URL and evidence location.
- Confirm the observed header, markup, rule, or text.
- Apply the smallest factual fix.
- Rerun the same tool and compare the rule ID, version, and evidence.