GEO AI
GEO Specification
The canonical implementation model for AI Search Optimization — four layers that define what it means for a site to be visible to AI search engines.
AI engines read differently
ChatGPT, Claude, Gemini, and Perplexity don't rank keywords. They ingest structured content — and ignore sites that don't provide it.
No standard existed
GEO Specification defines the exact signals AI crawlers expect: file formats, tag selectors, directive syntax, and schema types.
Compliance is measurable
The Analyzer evaluates every layer against the spec and produces a 0–100 score with actionable recommendations.
The four implementation layers
What GEO Specification defines
llms.txt
A plain-text file at your site root that gives AI models a structured map of your content — what you do, what pages exist, and how to navigate them.
Read specAI Metadata
Critical and supplemental HTML meta tags — title, description, Open Graph, Twitter Card — that AI crawlers read to understand page identity and context.
Read specCrawler Rules
Explicit allow directives in robots.txt for all major AI crawlers: GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and seven others.
Read specStructured Signals
JSON-LD schema markup (WebSite, Product, Article) that gives AI engines machine-readable facts about your content, products, and organization.
Read specAnalyzer
How the Analyzer uses the specification
Checks llms.txt
Verifies presence, content-type, and informative content at the site root.
Validates metadata
Confirms all four critical tags are present and non-empty.
Inspects crawler rules
Parses robots.txt and classifies each of the 10 known AI crawlers.
Evaluates structured signals
Detects JSON-LD blocks, microdata, RDFa, and sitemap fallback.
Scoring model
Every layer contributes 25%
The Analyzer evaluates each layer independently and assigns a status. The four category scores are averaged into a single 0–100 readiness score.
pass100 ptswarn65 ptsfail20 ptsnot_found10 ptsunknown40 ptsscore = round((llmsTxt + aiMetadata + crawlerRules + structuredSignals) / 4)
Definition
What does "GEO-ready" mean?
A site is GEO-ready when all four specification layers are implemented and passing the Analyzer. Each layer contributes equally to the overall score.
- llms.txt present, accessible, and informative
- All four critical metadata tags present
- robots.txt allows majority of AI crawlers
- At least one valid JSON-LD block present
Implementations
GEO AI packages implement this spec
Every package generates spec-compliant output. Pick the one that fits your stack.
Core
geo-ai-core
TypeScript engine. Zero dependencies. Works with any Node.js framework.
Next.js
geo-ai-next
Static file generation, middleware, and App Router route handler.
WooCommerce
geo-ai-woo
WordPress & WooCommerce plugin. Zero-config, WooCommerce-first.
Shopify
geo-ai-shopify
Shopify embedded app. Native metafields, App Proxy, multilingual.
Get started
Ready to implement the spec?
Check your current compliance score, read the full technical reference, or pick a package and start generating spec-compliant output today.