Make a workflow reproducible
A useful real example identifies the SDK version, input event, trace, result and failure boundary. A dashboard screenshot alone does not establish a measured outcome.
AI Agent Observability Website Review
Review the public evidence for your agent-observability product.
Use the scan for public HTML observations; review traces, integrations and outcome claims separately.
What RankFortune checks
- Raw HTML metadata and declared canonical
- Generic and matching-agent indexing directives
- Named search, training and user-fetch robots preferences
- One sitemap candidate, internal links and JSON-LD syntax
What you get back
- Observed issues and unknown checks, kept separate
- Sources, timestamps, scope and versioned rules
- Verification steps; actual citations and conversions remain not measured
Do not turn integration names into proof
Document the versions and behaviors actually supported. Distinguish tested functionality, planned integrations and user-supplied examples; do not imply this scan ran the SDK.
What happens after the scan
Review justified actions against the actual page goal. Add only truthful, applicable content or markup. The free scan does not crawl competitors, connect analytics or query target AI products; independent review and observations can proceed alongside it.
FAQ
Questions this audit answers
Does this scan measure actual AI visibility?
No. It is a public-page readiness review. The report marks actual AI mentions, citations and business outcomes as not measured.
Are FAQ, pricing and comparison pages mandatory?
No. They are only candidate content patterns. Their absence is not a failure or a reason to publish low-value variations.
Can I compare two scans?
Compare the same final URL and rule version, with the same intended scope. Preserve unknown states and confirm important changes with the appropriate external evidence.