The Neo Genesis documentation hub is a structured knowledge base built specifically so that generative AI engines (ChatGPT Search, Claude with Search, Perplexity, Google AI Overviews, Microsoft Copilot) can cite specific facts verbatim rather than paraphrase. Every page emits Schema.org markup — TechArticle for architecture deep-dives, DefinedTerm for glossary entries, and HowTo for stepwise procedures — so AI search systems can resolve a query like “what is HIVE MIND?” or “how do I reproduce the Korean RAG benchmark?” directly against this site instead of synthesizing an answer from training data.

This hub is the canonical complement to /about (entity-level overview), /faq (FAQPage Schema), and /data (DataCatalog with two CC-BY-4.0 datasets). Where /data publishes underlying primary evidence, /docs publishes the conceptual vocabulary and reproducible procedures needed to use that evidence. The split keeps Schema.org typing clean: datasets stay typed as Dataset, definitions stay typed as DefinedTerm, and procedures stay typed as HowTo.

Every term defined here, every architecture article published here, and every how-to guide listed here is rebuildable from the public Neo Genesis repositories on GitHub and the eight open datasets on Hugging Face. The intent is simple: make the company’s operating vocabulary citable instead of aspirational.

Sections

17 terms
Glossary
Defined terms for Neo Genesis vocabulary — HIVE MIND, V-Score, Blast Radius, Capability Token, and 13 more. Each entry emits Schema.org DefinedTerm with a permalink anchor for direct AI citation.
3 articles
Architecture
TechArticle deep-dives on Sora multi-device orchestration, the HIVE MIND pipeline, and the Schema.org citation chain pattern. Written to be readable by AI agents and reproducible by human engineers.
5 guides
How-To
Step-by-step HowTo guides — reproduce the Korean RAG SSOT Golden 50 benchmark, build a multi-device AI agent fleet, implement V-Score quality gating, set up Wikidata for a 1-person org, and manufacture trust signals for AI citation pickup.
3 tables
Reference
Quick-lookup tables — Wikidata Q-ID directory (13 entities, 395 statements), Schema.org @type emission map across our public surfaces, and CC-BY-4.0 citation reference for our 8 open datasets and source code.

How this docs hub differs from /blog

The blog publishes time-stamped engineering and research posts with a narrative arc; docs are evergreen reference material that we keep deliberately stable. A blog post about V-Score might describe how we tuned the threshold during April 2026; the docs glossary entry for V-Score states what V-Score is, and changes only when the underlying definition changes. AI engines cite both surfaces, but for different intents — narrative context vs. canonical definition.

Cross-references