Skip to content
AIWikis.org

AIWikis Information Architecture and Search Guide

AIWikis.org is a long-memory and dogfood surface. It should help a person, crawler, or coding agent know what a page contains before opening it.

Plain English

Use names that reveal the idea. "Part 1", "Page 1", "05 04 Notes", and hash-first URLs hide meaning. A good AIWikis route says what source, concept, guide, index, or evidence record it represents.

Use Search when you know words. Use concept and source-system routes when you know the domain. Use generated evidence pages only when you need provenance, hashes, bounded previews, or source-file history.

Technical Summary

AIWikis uses four public navigation layers:

  1. Orientation pages: start-here, what-is-an-LLM-wiki, governance, agent rules, and this guide.
  2. Search and index pages: /search/, /topics/, /source-map/, /files/, and /reports/.
  3. Concept pages: /concepts/uai/, /concepts/llm-wiki/, /concepts/ai-memory/, /concepts/provenance/, and related concept routes.
  4. Source-system hubs: /uaix/, /llmwikis/, /protocol5/, /justaniota/, and other source-memory guide routes.
  5. Evidence pages: generated source-file pages with source reference, hash, bounded preview, noindex guidance when appropriate, and links back to the owning source.

Every durable page should expose enough metadata for search and review:

  • title: human-readable page label, not a sequence label.
  • seo_title: optional title for browser, social, and search surfaces.
  • description and meta_description: public-safe summary, never raw JSON.
  • slug, canonical_url, and canonical: stable route identity.
  • page_type and primary_category: route grouping and browsing intent.
  • schema_type: JSON-LD type such as TechArticle, CollectionPage, or WebPage.
  • tags: terms for route index and agent retrieval.
  • source_site, source_status, trust_level, and noindex: authority and indexing boundaries.

Machine-readable route data is available at /wp-json/aiwikis/v1/search-index.

The human search page is available at /search/. It searches the same packaged route index that powers agent retrieval, so it works from the deployable Markdown package and does not depend on WordPress Page seeding.

Deep Spec

Naming Rules

Avoid:

  • Part 1
  • Page 1
  • Untitled
  • Misc
  • date-only titles such as 05 04 Intake
  • file-extension-first labels such as manifest json
  • hash-first public names

Prefer:

  • JustAnIota Protocol5 WordPress Memory Export Intake (2026-05-04)
  • LLMWikis Setup Wizard Guidance
  • AI Memory Source Archive Transfer
  • UAIX Project Handoff Evidence
  • Concept: LLM Wiki

Dates are evidence metadata, not the main topic. Put the date at the end when it distinguishes a source snapshot.

Route Hierarchy

Use this route model:

Route PatternUse
/start-here/Human and agent orientation.
/search/Human-facing search over route metadata, tags, aliases, descriptions, and previews.
/concepts/{concept}/Stable concept navigation.
/{source-site}/Source-system overview.
/{source-site}/source-memory-guide/How to read that source site's evidence.
/{source-site}/uai-system/Generated system-file index for a source site.
/{source-site}/uai-system/files/{semantic-source-reference}-{hash}/Generated file evidence.
/files/{source}/{semantic-source-reference}-{hash}/Global generated file evidence outside a source-specific hub.
/reports/Generated reports, inventories, and sync outcomes.

Generated evidence URLs may include hashes for traceability. The semantic portion must still name the source reference before the hash.

Indexing Boundaries

Source-file pages are useful retrieval evidence, but they should not compete with source-owned public pages. Use noindex: true and sitemap_include: false for raw generated source-file pages when they exist primarily for provenance. Keep hub, concept, guide, and index pages indexable when they provide public navigation value.

JSON Schema

Use content/schemas/aiwikis-page-metadata.schema.json for metadata validation. The schema exists to keep generated and hand-authored pages searchable, reviewable, and safe for public SEO surfaces.

Agent Rule

When repairing AIWikis pages, classify each page by intent before renaming. If a page title is vague, inspect the source reference, route, description, headings, and source site. Replace sequence labels with the smallest accurate concept or evidence label. Preserve aliases for old routes instead of breaking links.

Next Useful Routes

  • Start Here A task-first reading path for AIWikis.org, separating newcomer learning, source-memory lookup, maintainer workflow, and AI-agent retrieval.
  • Topic Index A tag-oriented index for LLM Wiki, AI memory, UAI, source governance, crawling, and retrieval topics.
  • Source Map AIWikis source-governed page for durable AI memory, evidence routing, and agent-readable retrieval.
  • Organization Memory Global organization namespace for shared AIWikis guidance and cross-site operating rules.