Site auditor
Enter a URL and get it scored against 39 checks drawn from the playbook checklist: metadata, structured data, crawling, llms.txt, performance and security headers, each with the evidence observed and the fix.
Public http and https URLs only. The auditor fetches the page, robots.txt, the sitemap, llms.txt and a 404 probe: at most 9 capped, timed-out requests. Nothing you enter is stored.
How to use it
- 01Enter your site's URLThe homepage is the best single page to audit because it carries the sitewide decisions. The auditor fetches it plus robots.txt, the sitemap, llms.txt and a 404 probe: at most 9 capped, timed-out requests.
- 02Read the score honestlyThe score counts only checks that passed or failed. Heuristic observations are marked improve, and checks whose precondition is missing are skipped rather than silently passed.
- 03Work the failures top to bottomGroups are ordered the way the playbook sequences the work: foundation first, because nothing above it can compensate for a broken base.
- 04Re-run after each deployEvery check here is deterministic, so a fix either moves the score or it does not.
Questions
What does the site auditor check?
The site auditor runs 39 checks against a URL: HTTPS and host canonicalisation, a real-404 probe, titles and descriptions, canonical links, Open Graph tags, JSON-LD structure and @id integrity, robots.txt and its AI crawler policy, sitemap lastmod differentiation, llms.txt, font hosting, and six security headers. 37 of the checks link to the corresponding entry in the full 112-check checklist.
Is the site auditor score a Lighthouse score?
The site auditor score is not a Lighthouse score. Lighthouse measures lab performance in a simulated browser; the site auditor measures implementation decisions visible in HTML, response headers and crawl files, which is the layer the playbook covers. The two are complementary.
Why does the site auditor skip some checks instead of failing them?
The site auditor skips a check when its precondition is missing: with no robots.txt there is nothing to declare a sitemap in, and with no JSON-LD there are no @id references to resolve. Skipped checks are excluded from the score, because one root cause should cost one failure, not five.
Does the site auditor store the URLs it audits?
The site auditor stores nothing. Each audit makes at most 9 rate-limited requests to the audited site, the responses are scored in memory, and the report is returned to the browser without being retained.
Where this comes from
This tool uses the same modules SEO for Solos sells in the Code Kit, so the output and the product cannot diverge. The reasoning behind it is in the playbook.