AI-only surface
purpose: expose stable crawlable records. human presentation: deprioritized.
distribution
format.html = /docs/<slug>/
format.md = /raw/<slug>.md
format.json = /api/pages/<slug>.json
site.index = /api/site.json
site.schema = /schema/knowledge-page.schema.json
site.llms = /llms.txt
site.map = /sitemap.xmlpages
| id | docs | raw | json |
|---|---|---|---|
| docs-home | /docs/ | /raw/docs-home.md | /api/pages/docs-home.json |
| information-architecture | /docs/information-architecture/ | /raw/information-architecture.md | /api/pages/information-architecture.json |
| url-conventions | /docs/url-conventions/ | /raw/url-conventions.md | /api/pages/url-conventions.json |
| page-template | /docs/page-template/ | /raw/page-template.md | /api/pages/page-template.json |
| json-schema | /docs/json-schema/ | /raw/json-schema.md | /api/pages/json-schema.json |
| retrieval-contract | /docs/retrieval-contract/ | /raw/retrieval-contract.md | /api/pages/retrieval-contract.json |
rules
- one topic -> one slug -> three synchronized outputs
- content in initial html response; no js dependency
- stable anchors, stable ids, stable metadata
- schema validation preferred over visual inspection