ARCHITECTURAL SPECIFICATION · 9 PILLARS

The 9 Deterministic Layers
Audited by AI Search Engines

ChatGPT, Perplexity, Google Gemini, and Claude do not merely count keyword occurrences. To place your brand in the recommendation set, they inspect technical and semantic consensus across these 9 pillars.

01 · DISCOVERY

Technical SEO & Crawlability

robots.txt permissions, clean sitemap hierarchy, HTTP status codes, canonical integrity, and AI crawler access (OAI-SearchBot, Claude-SearchBot, PerplexityBot). Unimpeded bot access is mandatory.

Target: HTTP 200 · Fail-closed SSRF Safeguard · 0 Loops
02 · MACHINE KNOWLEDGE

llms.txt & Markdown Surface

llms.txt v2 compliance, llms-full.txt, and clean Markdown alternates. Enables models to bypass DOM clutter (scripts, ads, cookie modals) and digest pure context with minimal token overhead.

Target: link rel="describedby" /llms.txt · <20KB Markdown
03 · GENERATIVE ENGINES

GEO (Generative Engine Optimization)

Semantic heading hierarchy, contextual density, and conceptual clarity enabling LLMs to segment and understand concepts without ambiguity.

Target: H1-H3 Semantic Tree · High Information Density
04 · DIRECT ANSWERS

AEO (Answer Engine Optimization)

Concise 50-word quotable direct-answer blocks and question-answering pairs, engineered for immediate extraction into search engine AI Overviews.

Target: <55-Word Quotable Snippet · Direct Definition
05 · TOKENOMICS

LLMO (Large Language Model Optimization)

Streamlined prose that respects model attention windows and KV-caches. Eliminates repetitive fluff to maximize the signal-to-token ratio.

Target: Low Perplexity · High Information-to-Token Ratio
06 · VECTOR RETRIEVAL

RAG (Retrieval-Augmented Generation)

Chunking-friendly layouts optimized for high cosine similarity and ColBERT MaxSim rankings within vector index retrieval stages.

Target: 256-512 Token Chunks · Cosine Similarity > 0.82
07 · AGENT READINESS

AAO (AI Agent Optimization)

Interoperability with autonomous agents (ChatGPT Operator, Claude Computer Use) via machine-readable endpoints, OpenAPI specifications, and A2A cards.

Target: OpenAPI Spec · A2A Agent Card · Form Accessibility
08 · ENTITY GRAPH

Schema.org & Entity Graph

Explicit JSON-LD @graph networks reconciling brand, founders, products, and services with canonical Wikidata QIDs and Google Knowledge Graph MIDs.

Target: 0 JSON-LD Errors · sameAs Knowledge Graph Anchors
09 · CONTENT TRUST

E-E-A-T & Trust Signals

Verifiable credentials, authorship accountability, primary source outbound links, and security hygiene (HTTPS/HSTS) mitigating LLM hallucination risks.

Target: Verifiable Author MID · RFC 3161 Timestamping

How Does Your Website Perform Across These 9 Pillars?

The HTML&HTML scanner audits your public domain against official specifications and exposes exactly what is broken with live evidence for free.

Scan Free Now →
AI SEARCH VISIBILITY → RECOMMENDATION OPPORTUNITY → CUSTOMER

Your customer asks AI ‘who should I choose?’ Is your website in the consideration set?

HTML&HTML prepares your website for visibility, citation eligibility and recommendation opportunity across AI search experiences. It shows measurable website-side blockers that can prevent discovery, understanding and source consideration.

01

BE DISCOVERABLE BY AI

robots.txt, sitemaps, canonicals, indexability and AI crawler access form the discovery foundation.

02

BE UNDERSTANDABLE

GEO, AEO, LLMO, entity graphs, schema and answer extractability reduce machine ambiguity.

03

BE SOURCE-READY

RAG/retrieval, original information, E-E-A-T, freshness and evidence support source eligibility.

04

TURN OPPORTUNITY INTO DEMAND

AAO, accessible journeys, intact links, measurable referrals and clear CTAs connect AI discovery to commercial action.

Recommendations, rankings, citations, traffic, customers and revenue are not guaranteed. HTML&HTML measures website-side technical and content blockers; it does not claim control over external AI systems.