SEO — Search & AI Engine Optimization

Production-ready SEO patterns for Next.js applications. Covers traditional SEO (Google), GEO (AI search engines), and implementation with actual code — not just recommendations.

This skill generates code, not audits. Every pattern includes production-ready Next.js App Router examples you can use directly.

Architecture Overview

[Metadata]
    ├── Static metadata (layout.tsx / page.tsx exports)
    ├── Dynamic metadata (generateMetadata with data fetching)
    ├── File-based (favicon.ico, opengraph-image.tsx, robots.ts, sitemap.ts)
    └── Title templates (%s | Site Name)
              ↓
[Structured Data]
    ├── JSON-LD via <script type="application/ld+json">
    ├── Schema.org types (Article, Product, Organization, BreadcrumbList)
    └── Rich results eligibility (Article, Product, Review, Video)
              ↓
[Crawlability]
    ├── robots.ts (Google + AI crawler control)
    ├── sitemap.ts (dynamic, with images and i18n)
    ├── Canonical URLs (alternates.canonical)
    └── Internal linking
              ↓
[Performance (Ranking Factor)]
    ├── LCP — next/image priority, Server Components
    ├── INP — useTransition, code splitting
    └── CLS — next/font, image dimensions
              ↓
[AI Visibility (GEO)]
    ├── llms.txt generation
    ├── Content structure for citation
    ├── AI crawler user-agent management
    └── Entity optimization

Quick Reference

Decision Matrix

"What metadata does this page need?"

"Should this page be indexed?"

Is the page valuable to search users?
├── YES (content, product, article)
│   ├── Is there a canonical version? → Set alternates.canonical
│   ├── Is it paginated? → Each page canonicals to itself. Link next with <a href>
│   └── Index it (default — no robots directive needed)
└── NO (auth, dashboard, admin, API, staging)
    └── Set robots: { index: false, follow: false }
        └── Do not Disallow that URL in robots.ts. Google must crawl it to see noindex.

SEO Anti-Myths (Google's Actual Guidance)

These are widely believed but wrong according to Google's official documentation:

What Google Actually Cares About

1. Unique, descriptive title tags per page
2. Concise meta descriptions that summarize page content
3. Canonical URLs to consolidate duplicate content
4. Structured data (JSON-LD) for rich results eligibility
5. Mobile-friendly design (responsive, touch-friendly)
6. Core Web Vitals — LCP <= 2.5s, INP <= 200ms, CLS <= 0.1
7. Descriptive alt text for images
8. Internal linking with descriptive anchor text
9. XML sitemaps for crawl efficiency
10. robots.txt for crawl control
11. HTTPS (ranking signal since 2014)

When Reviewing Code

• Every page has a unique <title> (via metadata export or generateMetadata)
• Every page has a meta description
• Root layout sets metadataBase for URL resolution
• Root layout uses title template: { template: '%s | Site Name', default: '...' }
• app/sitemap.ts exists and covers all public routes
• app/robots.ts exists with rules for Googlebot and AI crawlers
• Product/article pages have JSON-LD structured data
• Dynamic pages use generateMetadata with data memoization (cache())
• Images use next/image with proper alt text
• Fonts use next/font (prevents CLS)
• Auth/dashboard pages have robots: { index: false }
• Canonical URLs set for pages accessible at multiple URLs
• No <meta name="keywords"> tag (Google ignores it)
• OG images exist for shareable pages (social + messaging previews)

Performance follow-up sources and version limits: SOURCES.md. Read the affected resource for the task rather than treating field metrics or framework defaults as universal optimization requirements.