| name | seo-international |
| description | SEO internacional y local: hreflang, sitios multi-idioma/región, GEO (visibilidad en AI Overviews, ChatGPT, Perplexity, Claude). Cubre x-default, ccTLD, llms.txt, RSL 1.0. Usar cuando: "hreflang", "SEO internacional", "multi-idioma", "multi-región", "geo-targeting", "llms.txt", "AI search", "GEO", "aparecer en ChatGPT/Perplexity", "citas en IA", "país incorrecto en resultados", "x-default". "AI visibility", "generative engine optimization", "AI Mode", "cited by AI", "zero-click AI results", "AI search traffic", "local SEO", "Google Business Profile".
|
| allowed-tools | ["Read","Grep","Glob","Bash","WebFetch"] |
International SEO & AI Search Optimization (GEO)
Part 1: Hreflang & International SEO
When NOT to Use Hreflang
- Single-language, single-country site
- Machine-translated pages without real localization
- Content behind login or geo-blocking (must return 200 to Googlebot)
- No search presence in target markets
- Pagination/filter pages (only canonical pages)
Validation Checks
- Self-referencing: Every page MUST include hreflang pointing to itself. URL must exactly match canonical.
- Return tags: If A points to B, B MUST point back to A. Full bidirectional mesh.
- x-default: Required on every page with hreflang. Points to language selector or most universal version.
- Language codes (ISO 639-1):
| Correct | Wrong | Note |
|---|
ja | jp | jp = country, ja = language |
zh-Hans | zh-cn | Must use script subtag |
ko | kr | kr = country |
en-GB | en-uk | UK not valid ISO 3166-1 |
- Region codes: ISO 3166-1 Alpha-2, uppercase (
en-GB, not en-gb).
- Canonical alignment: Hreflang only on canonical URLs. Non-canonical = ignored.
- Protocol/URL consistency: All HTTPS, consistent www/trailing slash.
5 Most Common Errors
| Error | Fix |
|---|
| Missing return tags | Add matching B>A tags |
| Missing self-reference | Add self-referencing tag |
Wrong codes (jp, kr) | Use ISO 639-1 |
| Hreflang on non-canonical | Move to canonical URL only |
| x-default missing/wrong | Add to every page in set |
Implementation Methods
See references/hreflang-examples.md for code examples (HTML link tags, HTTP headers, XML sitemap) and generation process.
- HTML link tags: Best for <50 variants
- HTTP headers: Best for non-HTML (PDFs)
- XML sitemap: Best for large sites, cross-domain
Output: HREFLANG-AUDIT.md
Summary (pages, variants, issues, method), validation results per page, critical issues with fixes, generated code, GSC verification steps.
Part 2: AI Search / GEO Optimization (March 2026)
See references/geo-optimization.md for stats, brand signal correlations, AI crawler config, llms.txt, RSL 1.0, and platform-specific optimization.
GEO Analysis Criteria
| Dimension | Weight | Key Signals |
|---|
| Passage-level citability | 25% | 134-167 word answer blocks, direct answer first, specific facts, attribution |
| Structural readability | 20% | Question headings, short paragraphs, tables, lists, FAQ |
| Multi-modal content | 15% | Text + images/video/infographics/tools (156% higher citation rates) |
| Authority & brand signals | 20% | Author byline, pub dates, source citations, Wikipedia/Reddit/YouTube presence |
| Technical accessibility | 20% | SSR required (AI crawlers don't execute JS), robots.txt access, llms.txt |
Critical: Brand mentions correlate 3x more strongly with AI visibility than backlinks.
Output: GEO-ANALYSIS.md
- GEO Readiness Score: XX/100 (breakdown by dimension)
- Platform Scores table
- AI Crawler Access Status
- llms.txt Status
- Brand Mention Analysis
- Passage-Level Citability (3 best, 3 worst with rewrites)
- SSR Check
- RSL 1.0 Status
- Top 5 Highest-Impact Actions
- Content Reformatting Suggestions (before/after)
