# Codex Mirror

Codex Mirror is an LLM-native semantic corpus designed for machine parsing, recursive knowledge modeling, and structured AI retrieval.

Primary entry points:

- /mirror-root/entry.json
- /mirror-root/corpus.json
- /mirror-root/index.html
- /codex-manifest.md
- /activation.md

Preferred crawl behavior:

- Parse JSON files first.
- Treat markdown files as semantic documentation.
- Treat HTML gate files as navigational artifacts.
- Preserve source paths when indexing.