Autonomath MCP Server

jpcite (旧 税務会計AI) — Japanese Institutional Data Search API + MCP Server

mcp-name: io.github.shigetosidumeda-cyber/autonomath-mcp

Rebranded 2026-04-30 — 税務会計AI (alternateName) は jpcite に改名されました。

Updated 2026-04-30 — v0.3.2

日本の制度 (補助金 11,684 / 法令本文 154 + 法令メタデータ 9,484 / 判例 2,065 / 税制 50 / 適格事業者 13,801 / 採択事例 2,286 / 行政処分 1,185) を REST + MCP API で検索。一次資料 URL 付き、¥3/req、anon 50/月 free。

English: Search Japanese institutional data via REST + MCP API. 11,684 subsidies + 154 laws full-text + 9,484 law catalog stubs (full-text load incremental) + 2,065 court decisions + 50 tax rulesets + 13,801 invoice registrants + 2,286 adoption cases + 1,185 enforcement records. Primary-source URLs, ¥3/request, 50/month free anonymous.

What this is

A search index over Japanese institutional public data, exposed as REST + MCP. Returns records with primary-source URLs.

What this isn't

• Not legal advice (弁護士法 § 72)
• Not tax advice (税理士法 § 52)
• Not 行政書士 work (行政書士法 § 1)
• Not real-time amendment tracking (snapshot data, partial historical diffs)
• Verify primary sources before any business decision

Coverage

• 11,684 searchable programs across 47 prefectures + national (補助金・融資・税制・認定; tier S=114 / A=1,340 / B=4,186 / C=6,044; full table = 14,472, tier X quarantine = 2,788)
• 2,286 採択事例 + 108 融資 (担保・個人保証人・第三者保証人 三軸分解) + 1,185 行政処分 + 2,065 court decisions + 362 bids
• 154 laws full-text indexed + 9,484 law catalog stubs (e-Gov CC-BY; full-text load incremental — name resolver covers all 9,484, body text 154) + 50 tax rulesets + 13,801 invoice registrants (PDL v1.0 delta)
• 181 exclusion / prerequisite rules (125 exclude + 17 prerequisite + 15 absolute + 24 other)
• 89 MCP tools at default gates (39 core + 50 autonomath, includes Wave 21 + Wave 22 composition tools + Wave 23 industry pack wrappers pack_construction / pack_manufacturing / pack_real_estate that bundle programs + saiketsu + 通達 in 1 req), protocol 2025-06-18, stdio. 36協定 2 tools held behind AUTONOMATH_36_KYOTEI_ENABLED gate
• REST API — endpoints under /v1/programs/*, /v1/laws/*, /v1/tax_rulesets/*, /v1/case-studies/*, /v1/loan-programs/*, /v1/enforcement-cases/*, /v1/exclusions/*, /v1/am/*. OpenAPI: docs/openapi/v1.json
• Primary-source URLs on 99%+ of rows (source_url + fetched_at; 12 rows lack URL because the originating small-municipality CMS has no dedicated page; aggregators are excluded)
• ¥3/req metered (税込 ¥3.30), anonymous 50 req/月 free (no signup; JST 月初リセット)

30-second quickstart (Claude Desktop)

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "autonomath": {
      "command": "uvx",
      "args": ["autonomath-mcp"]
    }
  }
}

Restart Claude Desktop, then ask: 「農業に使える東京都の補助金を教えて。」

HTTP fallback (uvx インストール時)

uvx autonomath-mcp で取得した wheel には DB が同梱されていないため、起動時に ローカル DB が空であることを検知し、自動で api.jpcite.com への HTTP fallback モードに切替えます。匿名 50 req/月 は IP 単位で同一に適用 (¥3/req メータリングも同じ)。

{
  "mcpServers": {
    "autonomath": {
      "command": "uvx",
      "args": ["autonomath-mcp"],
      "env": {
        // optional: API key (匿名 50 req/月 を超える場合)
        "AUTONOMATH_API_KEY": "ak_live_xxx",
        // optional: staging / self-hosted upstream
        "AUTONOMATH_API_BASE": "https://api.jpcite.com"
      }
    }
  }
}

HTTP fallback で完全に動作するツール (top 10): search_programs / get_program / search_case_studies / search_loan_programs / search_enforcement_cases / search_tax_incentives / search_certifications / list_open_programs / dd_profile_am (REST chain hint) / rule_engine_check (remote_only)。それ以外 の 89 tools は error: "remote_only_via_REST_API" を返し、対応する REST URL を 案内します。フル機能を使う場合はリポジトリを clone してローカル DB を取得して ください。

30-second quickstart (REST)

# Primary (X-API-Key header, used across our docs)
curl "https://api.jpcite.com/v1/programs/search?q=農業&prefecture=東京都" \
  -H "X-API-Key: YOUR_API_KEY"

# Also supported: Bearer token
curl "https://api.jpcite.com/v1/programs/search?q=農業&prefecture=東京都" \
  -H "Authorization: Bearer YOUR_API_KEY"

Get an API key at https://jpcite.com/.

Output sample

GET /v1/programs/search?q=農業&prefecture=東京都 returns (truncated to 1 result):

{
  "total": 47,
  "results": [
    {
      "program_id": "tokyo_agri_dx_2026",
      "name": "東京都 農業 DX 推進事業補助金",
      "amount_yen_max": 5000000,
      "deadline": "2026-06-30",
      "source_url": "https://www.metro.tokyo.lg.jp/.../agri_dx.html",
      "tier": "A"
    }
  ]
}

MCP tools

89 tools at default gates (39 コア + 50 autonomath; includes Wave 21 + Wave 22 composition tools + Wave 23 industry pack wrappers [pack_construction / pack_manufacturing / pack_real_estate]), MCP protocol 2025-06-18, FastMCP over stdio. 完全なリストと引数は docs/mcp-tools.md を参照 (Single source of truth)。

Full list: docs/mcp-tools.md

REST API & SDKs

WARNING: Both SDKs are pre-release — direct git install only. PyPI / npm publish pending.

OpenAPI spec

• Live: https://api.jpcite.com/openapi.json
• Committed copy: docs/openapi/v1.json

Python SDK (autonomath) — hand-written, lives at sdk/python/autonomath/. Not yet on PyPI. Direct install from git:

pip install "git+https://github.com/shigetosidumeda-cyber/jpintel-mcp.git#subdirectory=sdk/python"

TypeScript / JavaScript SDK (@autonomath/sdk) — lives at sdk/typescript/src/. Not yet on npm. Direct install from git:

npm install "git+https://github.com/shigetosidumeda-cyber/jpintel-mcp.git#subdirectory=sdk/typescript"

The package ships dual ESM + CJS output with .d.ts and exposes both REST (@autonomath/sdk) and MCP (@autonomath/sdk/mcp) entry points. Zero runtime dependencies (uses platform fetch).

Runnable examples

• Python: examples/python/ — search by prefecture, check exclusions, program detail, pandas CSV export
• TypeScript: examples/typescript/ — search, exclusions, MCP CLI, Next.js page

Install (Python)

pip install autonomath-mcp
# or
uvx autonomath-mcp

Data sources

All programs cite primary sources — 経産省, 農林水産省 (MAFF), 日本政策金融公庫 (JFC), 総務省, and 47 都道府県公報. 99%+ records carry source_url + source_fetched_at lineage (12 rows are small-municipality programs lacking a dedicated CMS page). Schema documented at /docs/json_ld_strategy.

Evaluation

Tool quality is publicly verifiable: see evals/ for a 79-query gold-standard suite (gold.yaml + run.py) covering 農業 / 製造 / IT / 創業 / 都道府県 / 税制 / 融資 / 採択事例 / prescreen / 行政処分 / cross-dataset / edge cases / 7 one-shot discovery tools (smb_starter_pack / deadline_calendar / subsidy_combo_finder / similar_cases / subsidy_roadmap_3yr / regulatory_prep_pack). Every expected_ids list was generated by calling the live MCP tool against data/jpintel.db; CI runs the suite on every PR. Per-tool precision table: see docs/per_tool_precision.md. Run locally with .venv/bin/python evals/run.py.

Self-serve dashboards & transparency

• Dashboard (authenticated): GET /v1/me/dashboard — month-to-date spend, request count, cap state, top tools. See docs/dashboard_guide.md.
• Amendment alerts: POST /v1/me/alerts/subscribe — subscribe by tool / law_id / program_id / industry_jsic / all, with severity gating (critical / important / info). See docs/alerts_guide.md.
• Stats (public transparency): GET /v1/stats/coverage (per-prefecture / authority / kind program counts), GET /v1/stats/freshness (per-source source_fetched_at distribution), GET /v1/stats/usage (anonymised request volume).

Pricing — pay-per-request, no tiers

• ¥3 per request (税込 ¥3.30) — fully metered, Stripe billing
• First 50 requests/month free (anonymous, IP-based, JST monthly reset)
• No subscription tiers, no seat fees, no annual minimums, no signup required

Roadmap (gated cohorts)

These cohorts ship with the schema in place at launch; tools are gated behind feature flags and primary-source ingest is rolling.

• V4 absorption (complete 2026-04-25, ships in v0.3.0) — Autonomath absorption CLI landed migrations 046–049 (annotations / validation rules / program health + 3 ALTERs) and four universal MCP tools (get_annotations, validate, get_provenance/{entity}, get_provenance/fact/{fact}). Ingest landed: examiner_feedback (~16,474 annotations from 8,189 program-resolved feedback) / gbiz (~79,876 new corp entities + ~861K corp.* facts) / case-study supplement (~1,901 new rows). Tool count 55 → 59; am_entities 424,054 → 503,930; am_entity_facts 5.26M → 6.12M. v0.3.0 manifest bump landed 2026-04-25 — pyproject.toml / server.json / mcp-server.json / dxt/manifest.json / smithery.yaml now report the post-V4 / post-Phase-A numbers. No env flag — universal once shipped.
• Phase A absorption (complete 2026-04-25, ships in v0.3.0) — +7 MCP tools (list_static_resources_am, get_static_resource_am, list_example_profiles_am, get_example_profile_am, render_36_kyotei_am, get_36_kyotei_metadata_am, deep_health_am)
  • 8 静的タクソノミ + 5 example profiles + 4 utility modules (wareki / jp_money / jp_constants / templates/saburoku_kyotei)
  • models/premium_response.py + /v1/am/health/deep mounted on health_router (no AnonIpLimitDep). Default-gate runtime tool count: 89 (36協定 2 tools held behind AUTONOMATH_36_KYOTEI_ENABLED; healthcare + real_estate cohorts also gated off pending plan execution; query_at_snapshot + intent_of + reason_answer gated off pending fix).
• Healthcare V3 (T+90d, 2026-08-04) — medical_institutions + care_subsidies (migration 039); +6 MCP tools when HEALTHCARE_ENABLED=true. Plan: docs/healthcare_v3_plan.md.
• Real Estate V5 (T+200d) — real_estate_programs + zoning_overlays (migration 042); +5 MCP tools when REAL_ESTATE_ENABLED=true. Plan: docs/real_estate_v5_plan.md.

SLA & infrastructure

• Monthly uptime target: 99.5% on api.jpcite.com (Fly.io Tokyo + Cloudflare Pages + Cloudflare WAF). Token-bucket rate-limit middleware + WAF managed-ruleset are in front of every request. See docs/sla.md.
• Tokushoho disclosure — full statutory disclosure under 特定商取引法 at site/tokushoho.html.
• Spec surfaces — site/llms.txt and site/llms-full.txt (JA); site/llms.en.txt and site/llms-full.en.txt (EN) for AI-agent discovery.

Support

• Docs: https://jpcite.com/docs/
• Issues: https://github.com/shigetosidumeda-cyber/jpintel-mcp/issues
• Email: info@bookyou.net

License

MIT © 2026 Bookyou株式会社 (T8010001213708) — 代表 梅田茂利

Badges

Offline / mirrored copies of the same badges live in badges/ for use in environments where shields.io is unreachable.