MCUAPI/v1
mcuapi.up.railway.app
A free REST API for the Marvel Cinematic Universe: every movie and TV show, the characters and the actors who played them, and the chronology that orders it all — including the X-Men and Sony continuities the MCU keeps pulling in. No key, no signup, CORS open to any origin.
GET https://mcuapi.up.railway.app/api/v1/movies — no Authorization header
The catalogue isn't one universe either. Every title carries a continuity and an
Earth designation, and /timeline
returns them grouped and ordered by in-story chronology, so you can build a watch
order without maintaining one by hand. Hover a mark to read it; pick a branch to
load it in the console.
chronology_order within each branch, not by release date.
Snapshot embedded · refreshing from live API…
These calls hit mcuapi.up.railway.app
from your browser — same URL your app will use. Every
href in the response is
clickable, because the API is hypermedia-driven: you can crawl the whole graph
without hardcoding a single path.
No partial payloads and no second call to hydrate a title. Every field below is
on the resource by default — trim it with ?columns=
when you want less. Fields marked with a dot are nullable.
GET /api/v1/movies · /movies/{id}
GET /api/v1/tvshows · /tvshows/{id}
GET /api/v1/characters · /characters/{id}
played_by is per character record, and variant_of points a
character at the version they are a variant of — so a recast or a multiverse double
is a real edge in the graph, not a duplicate row you have to reconcile.
List endpoints wrap results as { data, total, page, limit, _links }
with first, prev, next and last
that preserve every filter you passed in.
?filter=column=value for a case-insensitive match,
?order=column,DESC to sort, and ?continuity= or
?is_mcu=false to scope to one branch of the multiverse.
Civilian names live on name and hero
names on alias, so the search hits both —
Captain America returns two different people. Every record carrying the term comes back,
including the variants each continuity casts separately, which is why
played_by
sits on the character and not on the title. Pick one and the page follows that record's own
_links.movies and
_links.tvshows to the titles it
appears in. No join table to maintain on your side.
variant_of; each one carries its own played_by.
Querying /characters…
# Every Phase 4 movie, newest first, three fields only curl "https://mcuapi.up.railway.app/api/v1/movies?filter=phase=4&order=release_date,DESC&columns=id,title,release_date" # The full chronological timeline for one branch curl "https://mcuapi.up.railway.app/api/v1/timeline?multiverse=Earth-10005"
const API = 'https://mcuapi.up.railway.app/api/v1'; // List endpoints return { data, total, page, limit, _links } const res = await fetch(`${API}/movies?filter=title=Spider&limit=5`); const page = await res.json(); // Follow the hypermedia link instead of building a path const cast = await fetch(page.data[0]._links.characters.href) .then(r => r.json());
import requests API = "https://mcuapi.up.railway.app/api/v1" # Walk the MCU in chronological order branches = requests.get(f"{API}/timeline?multiverse=Earth-616").json() for entry in branches[0]["entries"]: print(entry["chronology_order"], entry["title"], entry["type"])
# Optional — the API needs no client at all npm i mcuapi-client import { MCUAPI } from 'mcuapi-client'; const mcu = new MCUAPI(); // Fully typed. box_office is a string — Postgres bigint, // so the compiler stops you treating it as a number. const ironMan = await mcu.movies.get(1); // Walks every page for you by following _links.next for await (const c of mcu.characters.all()) { console.log(c.name, c.played_by); }
GET.
limit defaults to 10, capped at 100. Collections carry first, prev, next, last.
RateLimit-* headers and a 429 past it.
Cache-Control: public, max-age=3600 plus an ETag. Send If-None-Match for a 304.
Access-Control-Allow-Origin: * — call it straight from the browser.
Express, TypeScript, TypeORM and Postgres, organised as clean-architecture modules. If a title is missing, a cover is broken, or you want a field that isn't there yet, open an issue — the dataset is maintained by hand and every correction lands for everyone.