Menu
Audits & findings
The audit is where everything starts: SiteTR crawls the site, grades it, and hands back findings that each carry the evidence that produced them — never a bare number.
Quick & full
Two modes, one engine. Quick is the 30–90-second first-value scan — up to five pages, the expensive passes skipped. Full crawls up to 25 pages (10 by default) as mobile or desktop, checks links, and can add the AI-judgement pass and a live AI-bot probe for the GEO side.
Scores
Every audit produces an overall SEO score plus a GEO score — can AI assistants actually read, quote, and cite the site. Score history is stored per site, and a fall of ten points or more between re-audits raises an alert.
Findings & evidence
Findings are severity-graded — critical, high, medium, low — and each one names its evidence: the selector, header, request, or measurement that produced it, with a confidence level. The technical pass covers the classics (broken links, redirect chains, indexability, security headers) and the AI-era ones (bot access, llms.txt, structured data, answer-ready passages).
The fix plan
Findings roll up into a prioritized fix plan with impact ratings. Recommendations can ship as a patch or a GitHub pull request straight from the loop, and each carries a copy-paste verify step so the fix gets confirmed, not assumed.
Run one
From your agent with audit_site (read) or submit_audit (write, off unless the operator enables it), or the API above.
Exports
Issues export as CSV; the full numbers pack — issues, inspections, ranks — lands as a per-site XLSX workbook.