Flanaverse

An interactive map of Mike Flanagan's films and series: the actors who keep coming back, the props and easter eggs that tie the works together, and the authors he adapts. It's an unofficial fan project, live at https://flanaverse.elezea.com.

The site is static: Vite, TypeScript and D3, with no framework and no backend. Film data and images come from TMDB. Everything else (the connections, premises, actor profiles and trivia) is hand-curated from published interviews and articles, and every claim carries its sources and a confidence label.

Run it locally

You need Node 20.19+ or 22.12+ (Vite 8's minimum).

npm install
npm run dev          # http://localhost:5173

The generated TMDB data and images are committed, so you don't need a TMDB API key to run the site or to edit the curated content.

Command	What it does
npm run dev	Dev server on port 5173
npm run validate	Check the curated data: shapes, labels, and that every work, person, episode and image it mentions exists
npm run check	Type-check, then validate
npm run build	Type-check, validate, then build to dist/
npm run preview	Serve the built dist/

Contributing

Pull requests are welcome. If you've spotted a mistake but don't want to touch the code, suggest a correction instead: say where on the site it is and link a source, and I'll make the change.

The most useful pull requests are content fixes:

• Corrections. A wrong episode number, a misattributed quote, a prop that turns out to be a different prop. Point to the source that shows it.
• New connections. A prop, easter egg, callback, cameo, location or theme that the site is missing. Add it to data/curated/connections.json with at least one source.
• Better sources. Upgrading a reported or fan-theory entry to confirmed because Flanagan or someone from the production said it on the record.

Confidence labels mean specific things, and the site shows them to readers:

Label	Meaning
confirmed	Flanagan, a credited crew member, or an official production source has said it on the record (interview, social post, screen credit)
reported	Two or more independent outlets describe it, but nobody from the production has confirmed it
fan-theory	One outlet, a fan post, or an interpretation

Spoilers are fine. The site doesn't hide them.

Code fixes and new views are welcome too. For anything bigger than a bug fix, open an issue first so we can agree on the approach before you build it. docs/SPEC.md describes the routes, the shared visual vocabulary and the interaction rules that every view follows.

Before you open a PR:

1. Run npm run build and make sure it passes. CI runs the same build on every PR, and a PR can't merge until it passes.
2. Open the pages your change touches in npm run dev, in a narrow window as well as a wide one.
3. Keep the PR to one change. A batch of easter eggs is one change; an easter egg plus a CSS refactor is two.

If you use an AI coding agent, point it at AGENTS.md.

Where things live

Path	What
data/works.json	The list of works: slug, TMDB id, year, tier (core, early, written, upcoming)
data/curated/	Hand-curated content, typed in src/data/types.ts: connections.json (props, eggs, callbacks, references, cameos, locations, themes), works-meta.json (premises, roles, source material, trivia, episode notes), people-meta.json (repertory profiles), authors.json
research/	The research the curated content was written from, with sources and confidence labels
docs/SPEC.md	Routes, shared vocabulary, interaction rules, visual direction
src/data/generated/	TMDB output from npm run fetch plus the precomputed home map layout. Don't edit by hand
public/img/	TMDB posters, backdrops, stills and headshots. Don't edit by hand
src/data/model.ts	Joins generated and curated data; the only thing views read
src/views/	One module per route
src/components/	Shared chips, peek panel, "appears in" strip, search palette and trail
scripts/fetch-tmdb.mjs	TMDB pipeline
scripts/layout.mjs	Build-time force layout for the home map
scripts/validate.mjs	Curated data checks, run by npm run check, npm run build and CI

Refreshing TMDB data

Adding a work to data/works.json, or refreshing the cast and images, needs a TMDB API key. If you want a new work added and don't have a key, open an issue or a PR that only changes data/works.json, and I'll run the fetch.

echo 'TMDB_API_KEY=your-key' > .env                  # gitignored
npm run fetch                                        # fetch from TMDB, then recompute the layout
node scripts/fetch-tmdb.mjs --refresh && npm run layout   # same, ignoring the local cache
npm run layout                                       # recompute only the home map layout

npm run fetch reads data/works.json, caches raw API responses in data/cache/ (gitignored), downloads images into public/img/, and writes src/data/generated/works.json and people.json. TMDB's terms are non-commercial and ask that cached data be refreshed within six months.

Deploy

Only the maintainer deploys. The site runs as an assets-only Cloudflare Worker (wrangler.jsonc) on a custom domain, and merging to main does not deploy it.

npm run deploy       # build, then wrangler deploy
npm run og           # re-render public/og.png from og/og.html (needs npm run dev on :5173)

Licence

• Code is under the MIT License.
• Curated writing in data/curated/ and research/ is under CC BY 4.0.
• TMDB data and images in src/data/generated/ and public/img/ are not covered by either licence. They come from TMDB under the TMDB API terms, and the images belong to their rights holders.

By opening a pull request you agree that your contribution is released under the licence for the part of the repo it changes.

Credits

This product uses TMDB and the TMDB APIs but is not endorsed, certified, or otherwise approved by TMDB.

Easter-egg and prop research draws on Netflix Tudum, Collider, Screen Rant, CinemaBlend, Bloody Disgusting, Vulture, Deadline, Wikipedia and Flanagan's own posts. Each entry on the site lists its own sources.

Flanaverse is not affiliated with Mike Flanagan, Intrepid Pictures, Red Room Pictures, Netflix or Amazon.