Skip to main content

AutonoMath — Japanese public-program API + MCP server + npm SDK (66 tools): 13,578 programs + 2,286 case studies + 108 三軸分解 loans + 1,185 enforcement cases + 9,484 laws + 35 tax rulesets + 2,065 court decisions + 362 bids + 13,801 invoice registrants + 503,930 autonomath entities + 6.12M facts + 23,805 relations + 335,605 aliases, 181 exclusion/prerequisite rules, primary-source lineage. SLA 99.5% monthly uptime.

Project description

AutonoMath — Japanese Subsidy/Grant/Loan Database API + MCP Server

PyPI version License: MIT Python 3.11+ MCP 2025-06-18 CodeQL Made in Japan

日本の公的制度 (補助金・融資・税制・認定) 13,578 件 + 採択事例 2,286 + 融資 108 (担保・個人保証人・第三者保証人 三軸分解) + 行政処分 1,185 + 法令 9,484 + 判例 2,065 + 入札 362 + 適格事業者 13,801 を横断検索する REST + MCP API。AI エージェント・Claude Desktop・企業内 RAG 向け。

Why AutonoMath

  • 13,578 programs across 47 prefectures + national (補助金・融資・税制・認定)
  • 2,286 採択事例 + 108 融資 + 1,185 行政処分 + 2,065 court decisions + 362 bids — 単なる制度 DB ではなく「発見 + 併用可否判定 + 実績確認 + vendor vetting + 判例 + 入札」を 1 API
  • 181 exclusion / prerequisite rules for automatic 併給不可 + 前提条件 detection (125 exclude + 17 prerequisite + 15 absolute + 24 other kinds, 農業系 + 非農業系)
  • MCP native — 66 tools (38 コア: 制度検索・採択事例・融資・行政処分 + prescreen + upcoming_deadlines + 7 one-shot discovery (smb_starter_pack / subsidy_combo_finder / deadline_calendar / dd_profile_am / similar_cases / regulatory_prep_pack / subsidy_roadmap_3yr) + 拡張 [法令 e-Gov CC-BY 9,484 件・継続ロード中 / 税務ruleset インボイス+電帳法 35 件 live / 適格事業者 13,801 件 delta / 判例 2,065 件 / 入札 362 件 + cross-dataset glue] / 24 autonomath: 503,930 entities + 6.12M facts + 23,805 relations + 335,605 aliases の entity-fact DB + list_tax_sunset_alerts + V4 universal annotation/validation/provenance + Phase A static/example/template/health)、protocol 2025-06-18、stdio で Claude Desktop / Cursor / ChatGPT / Gemini から直呼び
  • REST API — 30+ endpoints: /v1/programs/*, /v1/laws/*, /v1/tax_rulesets/*, /v1/case-studies/*, /v1/loan-programs/*, /v1/enforcement-cases/*, /v1/exclusions/*, /v1/am/* (16 autonomath), customer self-serve /v1/me/* (cap / dashboard / usage_by_tool / billing_history / tool_recommendation / alerts / testimonials), transparency /v1/stats/{coverage,freshness,usage}, public /v1/testimonials. 完全な OpenAPI: docs/openapi/v1.json
  • REST fallback — curl / Python / Node / anywhere
  • Primary-source URLs on 99%+ of rows (source_url + fetched_at; 12件は小規模自治体 CMS 不在のため URL 未取得、aggregator は除外)
  • ¥3/req metered (税込 ¥3.30)、匿名 50 req/月 無料 (登録不要、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。」

30-second quickstart (REST)

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

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

Get an API key at https://autonomath.ai/.

MCP tools

66 tools (38 コア + 24 autonomath: 17 V1 + 4 V4 universal + 7 Phase A), 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)。

Group Tools
Programs search_programs, get_program, batch_get_programs, prescreen_programs, upcoming_deadlines
Case studies search_case_studies, get_case_study
Loans (3-axis) search_loan_programs, get_loan_program
Enforcement search_enforcement_cases, get_enforcement_case
Exclusion rules list_exclusion_rules, check_exclusions
Meta / vocabulary get_meta, enum_values
Laws (e-Gov, CC-BY) search_laws, get_law, list_law_revisions
Court decisions search_court_decisions, get_court_decision, find_precedents_by_statute
Bids (GEPS + 自治体) search_bids, get_bid, bid_eligible_for_profile
Tax rulesets (インボイス + 電帳法) search_tax_rules, get_tax_rule, evaluate_tax_applicability
Invoice registrants (PDL v1.0) search_invoice_registrants
Cross-dataset glue trace_program_to_law, find_cases_by_law, combined_compliance_check
AutonoMath entity-fact (autonomath.db) — V1 (17) search_tax_incentives, search_certifications, list_open_programs, enum_values_am, search_by_law, active_programs_at, related_programs, search_acceptance_stats_am, intent_of, reason_answer, get_am_tax_rule, search_gx_programs_am, search_loans_am, check_enforcement_am, search_mutual_plans_am, get_law_article_am, list_tax_sunset_alerts
AutonoMath V4 universal (4) get_annotations, validate, get_provenance, get_provenance_for_fact
AutonoMath Phase A (7) 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

REST API & SDKs

WARNING: SDKs are pre-release — published PyPI/npm packages coming after v0.1.0 tag.

OpenAPI spec

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

pip install "git+https://github.com/AutonoMath/autonomath-mcp.git#subdirectory=sdk/python"

TypeScript / JavaScript SDK (@autonomath/sdk) — lives at sdk/typescript/src/. Available on npm as well as direct git install:

# npm (preferred)
npm install @autonomath/sdk

# or direct from git
npm install "git+https://github.com/AutonoMath/autonomath-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

  • Free — 50 req/月 per IP (JST 月初 00:00 リセット、登録不要)
  • Paid — ¥3/req 税別 (税込 ¥3.30) metered (Stripe、外税表示)
  • No tiers. No seat fees. No annual minimums.

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 2026-04-25pyproject.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). Tool count 59 → 66.
  • 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.autonomath.ai (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 surfacessite/llms.txt and site/llms-full.txt (JA); site/llms.en.txt and site/llms-full.en.txt (EN) for AI-agent discovery.

Support

License

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

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

autonomath_mcp-0.3.0.tar.gz (675.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

autonomath_mcp-0.3.0-py3-none-any.whl (759.0 kB view details)

Uploaded Python 3

File details

Details for the file autonomath_mcp-0.3.0.tar.gz.

File metadata

  • Download URL: autonomath_mcp-0.3.0.tar.gz
  • Upload date:
  • Size: 675.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.12

File hashes

Hashes for autonomath_mcp-0.3.0.tar.gz
Algorithm Hash digest
SHA256 11df754dccd20352cc16dfa115359dc18e8cb5eb6f17093ad8ace18340bc7e51
MD5 3d813a1157f48bfe6ac3a9e548e98e66
BLAKE2b-256 85c91e0a0b3ea1140845c6d34b6f498f39bdadbf74c61ad091f58b3b1e0a04e2

See more details on using hashes here.

File details

Details for the file autonomath_mcp-0.3.0-py3-none-any.whl.

File metadata

  • Download URL: autonomath_mcp-0.3.0-py3-none-any.whl
  • Upload date:
  • Size: 759.0 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.12

File hashes

Hashes for autonomath_mcp-0.3.0-py3-none-any.whl
Algorithm Hash digest
SHA256 201a395c6eb20a1eb74ef76bec29de882155bcba69ac2d7afadd8bb83a37ffdb
MD5 469f5778e85a3e02a61f1c8d5ac05200
BLAKE2b-256 ced78816be086c0ea4ef85a31a8c8ffe3e006505569586355d2fa7400a5c43b8

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page