Next.js technical SEO checklist: 22 checks you can actually run

This checklist comes from real work: a 126-page site rebuild, then a bilingual restructure. Every item earned its place by breaking something.

Work through it in order. Don't skip the first eight — if the basics are wrong, everything after them is decoration.

---

1. Crawlability (8 checks)

1. Does the initial HTML contain body text?

The most common failure.

curl -s https://your-site.com/page | python3 -c "
import sys, re, html
raw = sys.stdin.read()
m = re.search(r'<main[^>]*>(.*?)</main>', raw, flags=re.S|re.I)
seg = m.group(1) if m else raw
t = re.sub(r'<(script|style)[^>]*>.*?</\1>', ' ', seg, flags=re.S|re.I)
t = re.sub(r'<[^>]+>', ' ', t)
print('body length:', len(re.sub(r'\s+', ' ', html.unescape(t)).strip()))
"

Pass: over 500 characters. I've seen 46 — the entire content was Loading….

Cause: client-side rendering. Crawlers don't execute JS on the first pass.

2. Is robots.txt blocking anything by accident?

curl -s https://your-site.com/robots.txt

Pass: no Disallow: /, sitemap URL correct.

Trap: a middleware allowlist that misses .txt — which silently tells AI crawlers the site is gone.

3. Is the sitemap complete and reachable?

curl -s -o /dev/null -w "%{http_code} %{size_download}\n" https://your-site.com/sitemap.xml
curl -s https://your-site.com/sitemap.xml | grep -c "<loc>"

Pass: 200, count matches actual pages.

Trap: lastmod set to new Date(). Every deploy marks all URLs as updated today; Google eventually distrusts the field and ignores it.

4. Are status codes correct?

for u in / /about /nonexistent-page; do
  printf "%-20s " "$u"
  curl -s -o /dev/null -w "%{http_code}\n" "https://your-site.com$u"
done

Pass: 200 for real pages, 404 for missing ones.

Trap: returning 410 for every unknown path. 410 means "permanently deleted" — a typo shouldn't tell search engines that, and you lose error monitoring signal.

5. Is canonical self-referencing?

curl -s https://your-site.com/page | grep -o '<link rel="canonical"[^>]*>'

Pass: points at itself, at the final URL (not a redirect source).

Trap: canonical says /services/en while that URL 308s to /en.

6. Any accidental noindex?

curl -s https://your-site.com/page | grep -o '<meta name="robots"[^>]*>'
curl -sI https://your-site.com/page | grep -i x-robots-tag

Pass: index, follow on content pages.

7. Single-hop redirects only

curl -sIL https://your-site.com/old-page | grep -E "^HTTP|^location"

Pass: no A → B → C chains.

8. Are static assets blocked?

curl -s -o /dev/null -w "%{http_code}\n" https://your-site.com/_next/static/chunks/main.js

Pass: 200. An over-broad middleware matcher that catches /_next/static produces a blank page.

---

2. Content structure (4 checks)

9. Exactly one H1

curl -s https://your-site.com/page | grep -c "<h1"

Pass: 1.

In a sample of 13 US small-business sites I audited, only 7 had exactly one — three had none at all.

10. Are H2s used for sections?

Pass: long pages broken into H2 sections.

Trap: <div class="text-2xl font-bold"> pretending to be a heading. Looks right, reads as nothing.

11. Unique titles

curl -s https://your-site.com/sitemap.xml | grep -o '<loc>[^<]*' | sed 's|<loc>||' | while read u; do
  echo "$(curl -s "$u" | grep -o '<title>[^<]*' | head -1) | $u"
done

Pass: no duplicates.

12. Meta description length

Pass: 60–155 characters. Enforce it in code with description.slice(0, 155).

---

3. Structured data (4 checks)

13. Is JSON-LD present?

curl -s https://your-site.com/page | grep -o '"@type":"[A-Za-z]*"' | sort | uniq -c

Page type	Expected types
Homepage	Organization, WebSite
Service page	Service, FAQPage
Article	TechArticle, BreadcrumbList
Product	Product, Offer

14. Are FAQs marked up?

A visible FAQ block without FAQPage schema is invisible to extraction. Add mainEntity with Question/Answer pairs.

15. Does Service reference provider by @id?

{ "@type": "Service", "provider": { "@id": "https://your-site.com/#organization" } }

Reference, don't redeclare — one entity, many mentions.

16. Are prices in the schema?

"offers": [{ "@type": "Offer", "name": "Starter", "price": "1200", "priceCurrency": "USD" }]

Pages with structured pricing get cited more readily when AI answers "how much does this cost".

---

4. Performance (4 checks)

17. Core Web Vitals

npx lighthouse https://your-site.com --output=json --output-path=lh.json --chrome-flags="--headless=new"

Metric	Target
LCP	< 2.5s
CLS	< 0.1
TBT	< 200ms

18. Modern image formats

UI screenshots should be WebP. Measured: four PNGs at 955 KB became 105 KB as WebP — 11%, and 6% on flat-colour screenshots.

19. Blocking scripts

A <head> script that rewrote a lang attribute pushed TBT from 60ms to 240ms. Removing it (static markup instead) brought it back to 60ms.

20. Fonts

An 8–10 MB CJK webfont destroys LCP. Use a system font stack or load a subset.

---

5. Internationalisation (2 checks)

21. Does html lang match the page language?

for u in / /en; do
  printf "%-6s " "$u"
  curl -s "https://your-site.com$u" | grep -o 'lang="[a-zA-Z-]*"' | head -1
done

Pass: / is zh-CN, /en is en.

This is the hardest one. In App Router, <html lang> can only be set once by the root layout. Per-route switching requires route groups:

app/(zh)/layout.tsx   <html lang="zh-CN">
app/(en)/layout.tsx   <html lang="en">

Patching it client-side doesn't work — crawlers see the wrong value.

22. Is hreflang bidirectional with x-default?

curl -s https://your-site.com/en | grep -o 'hrefLang="[^"]*"'

Pass: every language version lists all languages including itself, plus x-default.

Trap: page-level metadata replaces the layout's entire alternates object rather than merging. That's how /en lost its hreflang after a restructure.

---

Five-minute version

If you only have five minutes:

1. Body length on the homepage (check 1)
2. /robots.txt (check 2)
3. sitemap.xml count (check 3)
4. html lang (check 21)

If any of those fail, the rest doesn't matter.

---

Automate the basics

Checks 1, 2, 3, 9, 13 and 18 can be automated: free site audit tool — domain in, results in 30 seconds, no signup.

If it surfaces more than you want to fix yourself, the technical SEO service covers the full process and pricing.