Technical SEO Audit: links, redirects, canonicals, hreflang

Need a quick site audit for broken links, hreflang and meta tags? Give it URLs, get one audit row per page.

Crawl your own site (the start URLs you supply) and get one row per page plus a site summary. Obeys robots.txt, honours Crawl-delay, rate-limits (default 500 ms per host), stays on the start URLs' host, and identifies itself as NorthpineSEOAudit/1.0. Audit only sites you own or have permission to crawl.

Pay per result: about $8 per 1,000 audited pages ($0.008/page). Apify free credit covers small sites. No logins, no API keys.

What it checks

Per page: HTTP status, redirect chain, title and meta description (presence and length), H1 count, canonical (missing or pointing elsewhere), robots meta and X-Robots-Tag noindex, hreflang validity / self-reference / x-default, Open Graph and Twitter card, lang, viewport, images without alt, word count, internal/external link counts, an issue list with severity and a 0-100 score.

Site summary (key-value store, key SUMMARY): pages audited, average score, errors/warnings, redirect chains, duplicate titles and descriptions, broken internal and external links (with the pages that link to them), robots.txt, sitemap URLs and llms.txt presence. External links are checked with one HEAD request each and skipped if the other site's robots.txt disallows it. 401/403/429 results are reported as blocked-or-restricted, not broken.

Example input

{"startUrls":["https://example.com/"],"maxPages":50,"maxDepth":3,"checkExternalLinks":true,"delayMs":500}

Limits (honest)

Static HTML only: pages that need JavaScript to render their content or links are analysed as served. Not a Core Web Vitals or backlink tool. Scores are a rule-of-thumb summary, not a ranking guarantee. Not affiliated with any search engine.

Feedback and feature requests are welcome in the Actor's Issues tab. Built by Northpine Studio.

Dev

npm install && npm test; node test/run-local.cjs <url> [maxPages] for a local report; npx apify-cli push to deploy.

Caveat: project sites (e.g. user.github.io/project/)

robots.txt, sitemap and llms.txt are checked at the domain root, as crawlers do. A site served from a sub-path cannot publish files there, so "robots.txt not found" is expected and not a defect for those sites.

Example result

Real run on a small static site (8 pages, after fixing short titles/descriptions): average score 95, 0 errors, 1 warning, no broken links or duplicate titles/descriptions.

Billing and result rules

One row per distinct audited HTML page. Duplicate start URLs, redirect aliases of an already-audited page, non-HTML files and unreachable or error pages (404, 5xx, DNS failures) are not rows: they are listed in SUMMARY (unreachablePages, duplicateRedirects, skippedNonHtml, robotsBlockedUrls, invalidStartUrls). 429/5xx/network errors are retried with backoff. If the start URL redirects to another host (apex to www), that host is crawled too. Speed: expect roughly 1-8 s per page on cloud runs (politeness delay plus fetch time); for 1,000+ pages set a run timeout of several hours.