The Codex
The open database of comics, film, TV and anime
40,097 characters, teams, powers, creators, performers and productions, and 107,745 relationships between them — structured so an AI can use it without reading a single page, and open so anyone can build on it. Every row says where it came from.
Explore the web
Every record links to everything it touches. 5,309 team memberships · 9,657 creator credits · 19,806 screen credits · 3,494 rivalries · 7,578 family ties · 1,390 adaptations.
Characters
21,667 records
People
7,090 records
Works
1,999 records
Films
1,812 records
Episodes
1,318 records
Games
1,127 records
Anime
917 records
Manga
829 records
Comic series
807 records
Teams & organisations
728 records
TV shows
580 records
Powers & abilities
409 records
For AI agents and developers
The whole database as documented tables with keys — no scraping, no key, no rate limit. CC BY 4.0: use it, cite Comic Book Co.
Tables & schema
20 tables, every column documented
datapackage.json
Frictionless standard — tables, keys, licences
codex.sqlite.gz
everything in one relational file
REST API
static JSON + OpenAPI
llms.txt
how an AI agent should use the Codex
changelog.json
what changed, and when
Curated deep-dives
Hand-built and evidence-gated: every capability claim cites the issue it came from, and continuity is never blended.
Powers
37 cross-universe abilities
Characters
246 with 1,661 sourced feats
Commonly got wrong
191 corrections
Key issues
first appearances and milestones
Adaptation ledger
what screen work draws on which comic
Unadapted
great stories nobody has filmed, scored
Play
The Daily — Feat or Fake
five questions, new every day
Living views
pages that regenerate themselves
Where the data comes from
- Comic Book Co editorial — curated records, feats and corrections. CC BY 4.0.
- Wikidata — the structured spine: characters, creators, teams, relationships, screen credits and cross-database identifiers. CC0. Ingested by deterministic code; no AI writes a row.
- Deliberately not used — sites whose terms forbid automated collection, including fan wikis and entertainment news sites. A source's licence is checked before a single row goes in.
Last updated 2026-09-06
