Doomsday Protocol

An episode-exact Marvel watch tracker with deterministic progress, nightly slates, and an honest finish forecast.

---

Doomsday Protocol turns a large Marvel viewing order into a practical daily mission. It keeps three routes independent, tracks films and exact episodes, fits the next slate to the time available, and records where every progress observation came from.

Important

This is an unofficial, non-commercial fan project. Marvel properties, TMDB media/data, and the adapted MCU Watchlist compilation are not covered by the repository's MIT code license. See Third-party notices.

What it does

Capability	Behavior
Independent routes	Full journey (165), Doomsday path (72), and Essentials (60) retain separate progress and forecasts.
Exact missions	TV seasons advance episode by episode; films and specials remain atomic.
Nightly slate	Packs the next canonical items into a user-selected time budget.
Deadline model	Calculates remaining runtime, required daily pace, and projected finish date at 1× or 2× speed.
Deterministic progress	Manual, Jellyfin, Trakt, and import observations use normalized keys, timestamps, receipts, and replay-safe conflict rules.
Quiet tool surface	WebMCP exposes search, planning, provenance, preferences, and preview-before-import without cluttering the UI.

Stack

• Next.js 16 and React 19 through vinext
• TypeScript, Vite, and Tailwind CSS
• Cloudflare Workers and D1 through Drizzle ORM
• ChatGPT Sites authentication and hosting integration
• Node's built-in test runner

Run on another computer

Requirements: Git and Node.js 22.13.0 or newer.

git clone https://github.com/jerome-queck/doomsday-protocol.git
cd doomsday-protocol
npm ci
npm run dev

No private image bundle or API key is required to install, test, or build the project. The catalog is committed under data/; poster images load from the remote URLs recorded in that catalog. This is why a fresh clone looks like the same application without committing third-party poster files.

The public landing page works without an account. The private vault expects the authentication headers and D1 binding supplied by ChatGPT Sites. Local progress development uses the Workers/Vite integration configured in .openai/hosting.json; production account identity is intentionally not emulated or committed.

Verify a clone

npm run verify

That single command runs ESLint, a clean TypeScript check, the deterministic unit tests, and a production build. GitHub Actions runs the same command for every pull request and every push to main.

Repository map

app/                 Next.js routes, UI, authentication, and progress API
data/                Reproducible catalog and route membership snapshots
db/                  D1/Drizzle schema and binding access
drizzle/             Versioned database migrations
lib/                 Planner, progress contract, SQL, and store logic
public/              Original neutral project graphics only
scripts/             Catalog import and poster-URL refresh utilities
tests/               Deterministic unit tests

Data and media model

• data/watchlist.json is a versioned snapshot, not a live scrape at runtime.
• data/modes.json defines the three independent route memberships.
• Poster URLs point to TMDB or another identified rights-holder host; poster binaries are not copied into Git.
• Local screenshots, design references, and earlier franchise artwork are ignored. They remain on the creator's machine but are not needed by a clone.
• Database migrations and metadata are committed under drizzle/. Install a compatible Drizzle Kit release only when intentionally generating a new migration. node scripts/refresh-poster-urls.mjs refreshes poster references and performs network requests, so review its diff before committing.

Progress integrity

Every write uses a catalog-bound content key such as title:49 or episode:3:1. Imported activity is previewed before commit. Duplicate external events, stale observations, ambiguous matches, and equal-time conflicts fail closed instead of silently changing watched state.

Deployment

The canonical deployment runs on ChatGPT Sites. The repository contains the Sites project configuration but no credentials. Deploy from a ChatGPT Sites workspace authorized for the configured project; cloning the repository does not grant access to the existing production site or its D1 data.

Attribution and license

Original source code and documentation are licensed under the MIT License. The watch-order structure and editorial catalog are adapted from MCU Watchlist. Identifiers and poster references include data from TMDB.

This product uses the TMDB API but is not endorsed or certified by TMDB.

Third-party names, marks, media, data, and editorial compilations remain subject to their owners' terms and are excluded from MIT. Read THIRD_PARTY_NOTICES.md before redistribution or commercial use.