Japan Legal Corpus (e-Gov) MCP Server

A pay-per-call MCP server giving AI agents structured access to Japan's official legal corpus from e-Gov — Acts, Cabinet Orders, and Ministerial Ordinances. Article-level retrieval, revision diffs, and full source attribution in LLM-optimized JSON.

⚖️ Information retrieval only — not legal advice. This server returns text from the official e-Gov legal corpus. It does not interpret, summarize, or provide legal opinions. Always consult a qualified Japanese legal professional for advice on a specific matter.

What does this Actor do?

It exposes the Japanese government's legal corpus as five typed MCP tools so AI agents can look up statutes the same way they call any other tool. Behind the proxy is a Cloudflare Workers MCP server with KV-cached law data sourced from e-Gov; this Actor wraps that endpoint with Apify's pay-per-event billing, OAuth, and standby hosting so MCP clients can authenticate with their existing Apify token.

Why use this Actor?

• Compliance & due diligence — pull primary-source statute text into research workflows with citation URLs preserved
• Legal research agents — let an LLM resolve "which Article of the 労働基準法 covers overtime pay?" in one tool call
• Revision tracking — diff two revisions of the same law to surface changes between effective dates
• Multi-language — Japanese law text with English-friendly tool descriptions
• Sub-second — most queries complete inside the model's tool-call budget instead of stalling generation

Source-of-truth: every response carries the e-Gov publication metadata and the canonical laws.e-gov.go.jp URL.

Tools

Tool	What it does
search_law	Keyword search across Acts / Cabinet Orders / Ministerial Ordinances (Japanese or English keywords). Returns ranked law IDs and titles.
get_article	Fetch a specific article (条) from a law by ID + article number. Returns full text plus structural context (chapter / section).
list_categories	Enumerate law categories (e.g., 憲法, 民事, 刑事, 労働, 税). Useful for narrowing search scope.
get_law_metadata	Enactment date, last revision, promulgation number, competent ministry, current effective status.
compare_revisions	Diff two revisions of the same law/article and return added/removed/modified passages.

Performance

• p50: 196 ms warm (KV-cache hit on the upstream Worker)
• p95: < 1 s cold (cache miss — full e-Gov fetch + parse)

These are upstream-Worker numbers. Apify standby adds a ~50 ms proxy hop and an additional cold-start delay if the Actor has been idle.

Pricing

Pay Per Event: $0.018 per tool call. initialize, tools/list, and notifications are free. Only successful tools/call invocations are billed — JSON-RPC errors and isError: true responses are not charged.

Cost estimation

A typical legal research session runs ~10–30 tool calls (search → metadata → article → compare). At $0.018/call that's roughly $0.18–$0.54 per session. Bulk corpus extraction is also viable: 1,000 article fetches ≈ $18.

How to connect

Once the Actor is running in Standby mode, point any MCP-capable client at:

https://aulvem--egov-mcp-apify.apify.actor/mcp

with Authorization: Bearer <YOUR_APIFY_API_TOKEN> so Apify can meter and bill per call.

Example: Claude Desktop config

{
  "mcpServers": {
    "egov": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://aulvem--egov-mcp-apify.apify.actor/mcp",
        "--header",
        "Authorization: Bearer ${APIFY_API_TOKEN}"
      ],
      "env": { "APIFY_API_TOKEN": "<YOUR_APIFY_API_TOKEN>" }
    }
  }
}

Input

This Actor runs in Standby mode — there is no input form. Send JSON-RPC over HTTP to /mcp instead of starting a run with input.

Output

Tool responses are LLM-optimized JSON. Example for get_article:

{
  "lawId": "322AC0000000049",
  "lawTitle": "労働基準法",
  "lawTitleEn": "Labor Standards Act",
  "article": {
    "number": "37",
    "title": "時間外、休日及び深夜の割増賃金",
    "text": "使用者が、第三十三条又は前条第一項の規定により労働時間を延長し、又は休日に労働させた場合においては…",
    "chapter": "第四章 労働時間、休憩、休日及び年次有給休暇"
  },
  "revision": {
    "effectiveFrom": "2023-04-01",
    "promulgationNumber": "令和X年法律第Y号"
  },
  "source": {
    "publisher": "e-Gov",
    "url": "https://laws.e-gov.go.jp/law/322AC0000000049"
  }
}

You can download the dataset in various formats such as JSON, HTML, CSV, or Excel.

Data fields (article responses)

Field	Type	Description
lawId	string	Canonical e-Gov law ID
lawTitle	string	Official Japanese title
lawTitleEn	string	English title (when published by e-Gov)
article.number	string	Article number (条)
article.title	string	Article heading
article.text	string	Full article body
article.chapter	string	Containing chapter / section path
revision.effectiveFrom	string	ISO date the current text became effective
revision.promulgationNumber	string	Promulgation reference (法令番号)
source.url	string	Citation URL on laws.e-gov.go.jp

Companion server: Japan Statistics (e-Stat)

Pair this with the Japan Statistics (e-Stat) MCP Server for quantitative data — labor force, prices, population. Common workflows combine the two: pull the relevant statute via this Actor, then pull supporting time-series data from e-Stat.

FAQ

Is this legal advice? No. Information retrieval only. The text is sourced from e-Gov; interpretation is up to a qualified attorney.

Is the data current? Revisions are pulled from e-Gov. The upstream Worker refreshes its cache regularly, but for time-critical filings always verify against laws.e-gov.go.jp directly using the citation URL in the response.

Affiliation? Unofficial. Not affiliated with the Government of Japan, e-Gov, the Ministry of Justice, or Anthropic.

English coverage? e-Gov publishes English translations for selected major laws. lawTitleEn and English article text are returned when available; otherwise Japanese only.

Support / feedback? Open an issue on the Actor's Issues tab.