Technical SEO · AI website SEO
SEO for AI-Generated Websites: SSR, Canonical, Sitemap and Indexing
A deployed site may still give Google a JavaScript shell, wrong status codes, duplicate canonicals or a broken sitemap. Test what the published server actually returns.
Free implementation brief generator
Don't start from an empty chat. Choose the type of site or app you want, customise a detailed specification and copy it.
Start with 200, redirects and true 404s
Some single-page app deployments return 200 for every path, including non-existent URLs. Search engines may treat these as soft 404s. Check a real article, a permanently moved URL and a deliberately invalid path using curl -I. A canonical pointing to the homepage is not a replacement for a proper response code.
Test the actual production domain. CDN rules, hosting adapters and client-side routing may behave differently from local preview. Create a simple table of expected and observed statuses for the site's most important routes.
curl -I https://example.com/real-page curl -I https://example.com/old-page curl -I https://example.com/not-a-page
Initial HTML versus client-side hydration
Request the page with curl or inspect view-source. Look for the real article body, H1, canonical tag and links before React hydration. Rendering frameworks vary; the most useful question is whether Googlebot and people receive discoverable meaningful content.
JavaScript-only content is not automatically invisible to Google, but it adds rendering complexity. Build a server-delivered editorial structure even when interactive widgets are client-side. Check that hydration does not erase the initial text.
curl -Ls https://example.com/real-page | grep -Ei '<title|<h1|canonical|href=' | head -20
Canonical, robots, sitemap and international SEO
Each translation should carry its own self-referencing canonical and reciprocal hreflang references. A translated English page should not canonicalise to a Ukrainian homepage. Sitemaps should contain only canonical indexable URLs, with meaningful lastmod values.
Robots.txt controls crawling but doesn't safeguard private data. Check meta robots and X-Robots-Tag after deployment, and make sure staging copies do not compete with production. Examine Search Console's actual exclusion reasons before retrying indexing.
Structured data and real Core Web Vitals
Article data should identify a truthful author and dates. Product schema must reflect real visible price and stock. Fabricated Review or AggregateRating properties should be removed. Google's structured-data support is narrower than the whole Schema.org vocabulary, and valid markup doesn't guarantee a rich result.
Review LCP, INP and CLS on actual devices. Compress unnecessary hero assets, reserve space for media and avoid loading heavy scripts just to animate decoration. The most useful SEO optimisation often also makes the site easier for customers to use.
| Check | Where | Expected |
| HTTP | curl -I | 200/301/404 as appropriate |
| SSR | curl -L | Article and links included |
| Canonical | HTML head | Self-referencing URL |
| Hreflang | Head/sitemap | Reciprocal languages |
| Indexing | Search Console | Actual evidence |
A practical first-build scenario
One-URL technical SEO QA
- 01
Fetch headers and initial response HTML.
- 02
Check title, H1, canonical, robots and hreflang.
- 03
Verify sitemap inclusion and discoverable internal links.
- 04
Inspect the final published URL through Search Console.
Definition of done: Every pass or fail is based on the real server response, not a builder's automated claim.
Field workbook · outcomes, not theory
Audit an AI website's actual SEO HTML without executing JavaScript
A page looking correct in a browser does not prove its crawler HTML is complete. Save a raw response, check HTTP status, initial H1 and body, canonical, robots, sitemap and non-existent page handling before touching meta tags.
Reproducible HTTP inspection commands
Replace example.com and the path. This is synthetic HTTP testing, not an actual crawl by Google or Search Console Request Indexing.
DOMAIN='https://example.com'
PATH='/services/example'
curl -sSIL --max-time 20 "$DOMAIN$PATH"
curl -sSL --max-time 20 "$DOMAIN$PATH" -o page.html
grep -Ei '<title|<h1|canonical|noindex|<main' page.html | head -40
curl -sSL "$DOMAIN/sitemap.xml" -o sitemap.xml
curl -sSL "$DOMAIN/robots.txt"
curl -sS -o /dev/null -w '%{http_code}\n' "$DOMAIN/test-404"Do this in order
- 01
Inventory every canonical URL
Merge the sitemap list with internal HTML links; identify missing/duplicate/orphaned pages.
Evidence: CSV with each URL and response.
- 02
Inspect first HTML, not hydrated DOM
Fetch the actual final URL with curl and record H1, meaningful paragraphs, title and canonical.
Evidence: Proof that text is visible without JS.
- 03
Repair routing and HTTP errors
Use leaf index routes or an Outlet where needed, correct 301/404 and canonical mismatches.
Evidence: Direct URL and navigation both render the same content.
- 04
Compare with official GSC data
Inspect actual Google-known status in URL Inspection and group problems by cause; do not confuse synthetic User-Agent requests with indexing.
Evidence: Two separate reports: HTTP crawl and GSC findings.
Troubleshooting the failed result
Click changes URL but chapter stays unchanged
Likely cause: Parent route omits nested Outlet or incorrectly catches child.
Action: Fix layout/leaf route and add browser navigation test.
curl gets an empty app shell
Likely cause: Content relies on client hydration.
Action: Serve meaningful SSR/static HTML.
Crawled, not indexed remains
Likely cause: Technical duplication, weak unique value or other selection causes.
Action: Use GSC diagnosis, consolidate and improve pages as needed.
Acceptance checklist
0/6Check only outcomes you've verified yourself. Your progress stays in the current page session and is reset on reload.
Verify every criterion before launch.
Passing crawlability checks does not guarantee inclusion in Google's search index.
Questions before you begin
Is SSR required for Google indexing?
Google can render JavaScript, but meaningful initial HTML simplifies discovery and reliability.
Does valid schema guarantee a rich result?
No. Eligibility and display remain Google's decisions.
Why isn't my site indexed?
Check status codes, robots, canonical, content usefulness and Search Console evidence.
Next: validate your result
After building an app, test its real Google search eligibility and measurement. These are independent acceptance guides, not more AI prompts.