PokéPC Dataset

Static JSON data for Pokémon, games, Pokédexes, box presets, and related PokéPC metadata.

The package also includes TypeScript helpers, Zod schemas, and an OpenAPI description for serving the dataset as a static JSON API.

Install

pnpm add @pokepc/dataset

Example imports:

import bulbasaur from '@pokepc/dataset/data/pokemon/bulbasaur'
import { pokemonSchema } from '@pokepc/dataset/lib/schemas'

const pokemon = pokemonSchema.parse(bulbasaur)

JSON imports depend on your runtime or bundler configuration. In this repository, the raw files are always available under data/.

Static API

The OpenAPI docs and static JSON API are hosted on GitHub Pages:

• API client (Swagger UI) https://pokepc.github.io/dataset/
• OpenAPI spec: https://pokepc.github.io/dataset/openapi.json

| URL path | Dataset source | | ------------------ | ----------------------------------------- | | /dataset/ | Latest default-branch build (development) | | /dataset/latest/ | Latest stable SemVer tag | | /dataset/v6/ | Latest stable v6 tag | | /dataset/v7/ | Latest stable v7 tag |

Each path serves its own openapi.json, data/, and data-next/. Every spec lists all deployed servers, with its own server selected first in Swagger UI. The documentation links open each version's matching schema; changing the server dropdown only changes the request destination. versions.json records the deployed refs and commits.

Retained major paths are configured in pages-versions.json. Add a major after its first stable tag exists to expose /vN/; removing one removes that path on the next deployment. Root and /latest/ are always included. Tags may use 6.9.1 or v6.9.1 spelling; prereleases are excluded and versions are compared numerically. Conflicting equal-precedence tags, missing configured releases, or tag/package version mismatches fail the build.

Pages rebuilds for default-branch source/configuration changes and after a successful Publish npm Packages workflow. To refresh independently of npm, run Deploy to GitHub Pages manually from Actions. Every deployment replaces the site with one complete artifact, preserving all configured major paths. Historical builds use their own code and frozen dependency lockfiles.

Use Cases of the openapi.json spec

Apart from being able to use it in your code with validators or generators, when combined with AI agents, it unlocks many extras:

1. Tool calling / agent usage: The LLM can know what endpoints exist and how to call them.

2. Code generation: It can generate typed clients, fetch wrappers, SDK helpers, Zod schemas, etc.

3. Documentation Q&A: Users can ask natural-language questions and the LLM can map them to the right endpoint/schema.

4. Safer answers: Enums like language IDs, generation IDs, region IDs, ribbon IDs, etc. reduce hallucinations.

5. Dataset exploration: The LLM can inspect available resources without manually reading every file.

Data Layout

data/
  abilities.json
  codes/
  games/
  indices/
  metadata/
  pokedexes/
  pokemon/
  boxpresets/
    classic/
    modern/

Root JSON files are collection files. games/, pokedexes/, and pokemon/ contain one JSON file per entity. indices/ controls the order of those per-entity files. codes/ holds append-only numeric codes for Pokémon, ribbons, marks and moves; read the code map rules before adding, renaming or removing any of them.

The game lifecycle reference defines release, digital delisting and required-service shutdown dates, including announced future closures.

The Pokémon availability field reference defines acquisition, storage, shiny availability, and related form fields for both maintainers and agents. The evolution reference and form-transition reference explain the method arrays, typed conditions, and translation-ready enum exports. The ability-history audit defines legacyAbilities and lists released historical ability changes, with unreleased/unused assignments documented as exclusions.

Contributing

Requirements: Node.js 24 and pnpm 11.

pnpm install
pnpm format:check
pnpm typecheck
pnpm test
pnpm build

When changing data, run the tests before opening a PR. When changing the static API docs, run pnpm build:pages and check the local Swagger UI with pnpm dev:openapi.

pnpm test blocks external network requests, including in Node CLI subprocesses. Upstream tests must use fixtures, mocked responses, or loopback test servers; they must not crawl live services.

Build a GitHub Pages-ready artifact:

pnpm build:pages

This writes:

dist-pages/
  index.html
  openapi.json
  data/
  data-next/

Build all published versions from the remote default branch and tags (requires Git, Node.js 24+, Bun, pnpm, and network access). This uses the current checkout's deployment configuration and Swagger shell while leaving its source files and index unchanged:

pnpm build:pages:versions

The aggregate output adds latest/, configured vN/ directories, and versions.json to dist-pages/. Sources pointing to the same commit are built once. All sources must build successfully before the output is replaced. CI pins the root source to its tested checkout with --root-sha=<full-commit-sha> so a concurrent push cannot mix the checked configuration with a newer default-branch build.

To build and preview the aggregate site locally under the same project prefix:

pnpm build:pages:versions --base-url=http://localhost:4173/dataset/
pnpm dev:pages:versions

Open http://localhost:4173/dataset/. This preview serves the assembled files without rebuilding them; every server option targets the local preview. Keep the build URL's port aligned with the preview's PORT or --port setting. A deployment build uses the public package.json homepage as its base URL by default.

Preview the Swagger UI locally:

pnpm dev:openapi

The dev server runs at http://localhost:4173/ by default. Override the port with PORT=4174 pnpm dev:openapi or node src/openapi/dev.ts --port=4174.

Dataset editor

Edit this checkout's data/ with the local maintainer app:

pnpm dev:editor

The editor runs on http://127.0.0.1:3003 without external-directory configuration. pnpm dev:openapi starts only the OpenAPI preview, and pnpm dev runs both in parallel. See the editor guide for editing rules, builds and disposable browser tests.

Pokémon availability lookup

Game records include PokéAPI version and version-group IDs. See the game ID mapping guide for the schema and one-off population script.

Inspect the two Bulbapedia availability lists using a dataset Pokémon ID or nid:

pnpm pokemon:availability pikachu
pnpm --silent pokemon:availability 0026-alola --json
pnpm pokemon:availability pikachu --patch
pnpm pokemon:availability:all --dry-run

The command prints a terminal table with one row per game or candidate Pokémon availability fields. It preserves existing values where the source is inconclusive, including storableIn, and reports warnings. With --patch, it updates and formats the selected Pokémon file and prints a summary of added and removed games instead of the table or JSON. The CLI and editor share deterministic parsers for the main and GO availability lists; no AI or API key is required. See the CLI guide for saved HTML input, classification rules, and limitations.

Run pnpm pokemon:availability:all to review every Pokémon interactively. Each iteration shows the changes summary and accepts p to patch or s to skip. Press Ctrl+C to stop.

Credits

This dataset uses public data from:

• Pokémon Showdown
• PokéAPI
• Project Pokémon
• Serebii.net
• Bulbapedia

Thanks to everyone who collects and maintains public Pokémon game data.