Technical SEO Audit Checklist 2026: 34 Checks That Matter

Technical SEO is the plumbing. Nobody notices it until a pipe bursts, and then it is the only thing that matters. A template update adds noindex to 400 product pages, a migration leaves every canonical pointing at the old domain, a framework upgrade moves the H1 into JavaScript. Rankings fall and the content team gets blamed.

This technical SEO audit checklist has 34 checks in six groups. Each has a pass condition you can verify, so the audit produces facts rather than opinions. It is the technical half of the broader SEO audit checklist, which adds on-page relevance and content.

Why a technical audit comes first

On-page and content work assumes the page can be crawled, rendered and indexed. If it can't, rewriting the title does nothing. That is why every technical audit starts with access and works outwards.

Technical problems also scale differently from content problems. A weak article affects one page. A broken template affects every page built on it, often thousands. In Seotific's scoring, most critical-tier findings are technical: noindex, canonical errors, pages not indexed. They cost 10 points each, against 2.5 for a medium-tier content issue.

Two 2026 changes matter here. Google's Search Console now reports generative AI impressions, so pages that fail technical checks are also missing from AI Overviews and AI Mode. And because most AI crawlers do not execute JavaScript, a rendering problem that Google tolerates can still hide your content from ChatGPT and Claude.

How to run a technical SEO audit in 9 steps

1. Open Search Console's Page indexing report and note the indexed count, the not-indexed count and the top three reasons.
2. Crawl the site with a crawler that can render JavaScript, starting from the homepage and also from the XML sitemap.
3. Compare the two crawls: URLs in the sitemap but not found by links are orphans; URLs found by links but not in the sitemap need a decision.
4. Filter the crawl for non-200 status codes, redirect chains and internal links pointing to redirects or errors.
5. Filter for indexability: noindex in meta or header, canonicals pointing elsewhere, robots.txt blocks.
6. Spot-check ten important templates with JavaScript disabled to confirm the title, H1, canonical and body text are in the raw HTML.
7. Pull Core Web Vitals field data for the main templates from PageSpeed Insights or the Core Web Vitals report.
8. Check HTTPS, host and trailing-slash consistency: every variant should 301 to one canonical URL.
9. Record each failure with its template, the number of affected URLs and the fix, then rank them by impact.

The 34-check technical SEO audit checklist

Crawling (checks 1 to 8)

1. robots.txt returns 200 and is under 500 KiB, the most Google reads.
2. Important paths are not disallowed. Search for Disallow: rules that match product, category or blog URLs.
3. The sitemap is listed in robots.txt with a full URL.
4. The sitemap lists only 200, canonical, indexable URLs. No redirects, no noindex, no parameter duplicates. Maximum 50,000 URLs or 50 MB per file.
5. No redirect chains in internal links. Links point at the final URL; one 301 at most.
6. No internal links to 4xx pages. Fix or remove them.
7. Server response time stays under about 800 ms time to first byte on key templates.
8. Search Console crawl stats show no host availability problems (DNS, server connectivity, robots.txt fetch).

Indexing (checks 9 to 16)

9. No unintended noindex in the meta robots tag or the X-Robots-Tag header.
10. Canonical tags are self-referencing on pages you want indexed.
11. Canonical URLs return 200 without redirecting.
12. One URL per page. HTTP to HTTPS, www to non-www (or the reverse) and trailing-slash variants all 301 to one version.
13. Filter and sort parameters are canonicalised to the clean URL or kept out of the crawl.
14. No soft 404s: empty search results, out-of-stock pages and placeholder pages either return 404 or carry real content.
15. Page indexing reasons reviewed in Search Console, especially crawled, currently not indexed and discovered, currently not indexed.
16. Indexed count is close to the sitemap count. A gap of more than 20% needs explaining.

Rendering (checks 17 to 21)

17. Title, H1 and canonical are in the raw HTML, not added by JavaScript.
18. Main body text is in the raw HTML or reliably server-rendered.
19. Internal links are real <a href> elements, not click handlers on <div> elements.
20. URL Inspection's rendered HTML matches what users see.
21. CSS and JavaScript needed for rendering are not blocked by robots.txt.

Site structure and internal links (checks 22 to 27)

22. Important pages are three clicks or fewer from the homepage.
23. No orphan pages: every indexable page has at least one internal link. See orphan pages and SEO.
24. Breadcrumbs on deep pages, with BreadcrumbList structured data.
25. Readable URLs: lowercase, hyphen-separated, no session IDs.
26. Pagination uses plain links so page 2 onwards can be crawled.
27. No duplicate titles or H1s across different pages.

Speed and mobile (checks 28 to 31)

28. Responsive viewport tag on every template.
29. LCP 2.5 seconds or less at the 75th percentile of real visits.
30. INP 200 milliseconds or less.
31. CLS 0.1 or less.

Security, international and structured data (checks 32 to 34)

32. HTTPS everywhere, with every HTTP URL redirecting and no mixed content.
33. hreflang is reciprocal and complete, only if the site has several languages or regions.
34. No structured data errors in Search Console's enhancement reports or the Rich Results Test.

Reading Search Console's indexing statuses

The Page indexing report is the closest thing to Google's own audit of your site. Each "Why pages aren't indexed" reason points to a different technical check:

Two of these come up often enough to have their own guides: duplicate without user-selected canonical and canonical tag points to another page.

The technical issue most audits miss: JavaScript rendering

Classic crawlers read raw HTML; Google renders the page in a headless Chrome and indexes the result. When the two differ, a crawler can report "H1 missing" on a page Google sees fine, or report everything present on a page whose content only appears after a script loads.

The risk is not that Google can't render JavaScript. It usually can, but rendering is queued and can lag behind crawling, and errors in scripts can leave the rendered page empty. The bigger risk in 2026 is everyone else. Most AI crawlers do not execute JavaScript, so a React or Vue page that ships an empty shell is invisible to ChatGPT search, Claude and Perplexity even if Google indexes it.

The test is simple: load the page with JavaScript disabled, or compare "view source" with the rendered DOM, and confirm the title, H1, canonical, main text and internal links are in the HTML the server sends. The full method, with the frameworks that get it wrong most often, is in the JavaScript SEO rendering check guide, and the framework-specific fixes for Next.js are in the Next.js SEO checklist.

Technical audit tools compared

For most sites the practical stack is Search Console plus one crawler. Seotific covers the crawler role with the fix order built in, and its re-audits show which issues are new, fixed or still open after each release. For agencies, the SEO audit checklist template turns the same checks into a client-facing sheet, and the WordPress SEO audit checklist adds the platform-specific traps.

Frequently asked questions