Home /

Everything on this site is built from structured, human-verified data. We publish that data openly, so that apps, researchers, other cities — and AI assistants — can use it directly instead of scraping pages.

JSON API

No authentication, no rate limits, CORS enabled. Please cache responses.

• /api/procedures.json — every procedure as structured steps, with deadlines and human-verified official links
• /api/wards.json — all 23 wards: offices and foreign residents' desks (address, phone, hours, languages)
• /api/glossary.json — Japanese terms from official documents, explained in 5 languages
• /api/emergency.json — 119, #7119 and multilingual helplines
• /api/meta.json — freshness check (generated date, oldest verification)
• OpenAPI spec · JSON Schema

Every record carries a last_verified date — the date a human opened and checked it.

Markdown mirrors & llms.txt

Every page on this site has a plain-Markdown twin: add index.md to any page URL (for example /moving-in/resident-registration/index.md). Site indexes for AI live at /llms.txt and /llms-full.txt.

MCP server

AI agents can call this data as tools instead of searching the web:

{
  "mcpServers": {
    "official-tokyo": {
      "command": "npx",
      "args": ["-y", "@dror-jp/official-tokyo-mcp"]
    }
  }
}

Package: @dror-jp/official-tokyo-mcp — dependency-free, five tools (procedures, wards, glossary, emergency contacts).

Why it never goes stale

The JSON, the Markdown mirrors and the HTML pages are all generated from the same source files at build time. There is no separate database to fall out of sync — if a page is corrected, the API is corrected in the same moment.

License

Data: CC BY 4.0 — attribution official.tokyo.jp (DroR Corporation). Code: MIT, open source on GitHub. The data structure is deliberately not Tokyo-specific — if you want to build this for another city, fork it. That is the point.