Summary
A focused collection of Japan-specific data endpoints wrapped in an MCP interface. You get weather lookups via Open-Meteo, a railway archive with 1,800+ defunct Hokkaido stations, Japanese food composition tables from MEXT, SPDX license data, and HTTP header registries. Most queries are free for human use. AI crawler access to premium endpoints triggers x402 micropayments on Base (USDC), with weather calls priced at $0.001 and archived railway PDFs at $0.01 each. The server runs on Cloudflare Workers and exposes structured JSON through streamable HTTP transport. Useful if you're building agents that need Japanese public datasets or want to experiment with pay-per-call API models without managing API keys.
Keep your Mac awake while Claude Code and 40+ AI agents run. Sleeps when they're idle.
One time payment $9 →
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
On Capafy, your Skill runs online 24/7 as an agent product, and you get paid every time someone uses it.
Start earning →
Micro Data API Factory
Public-source structured data APIs for AI agents. x402 micropayments.
Live: https://micro-data-api-factory.kasanegi123.workers.dev
Datasets
| Dataset | Items | Description |
|---|---|---|
| Japanese Food Composition | 2,538 | MEXT official nutrition data — 18 food groups, 100g basis |
| SPDX License List | 727 | Complete OSS license catalog with OSI/FSF status |
| HTTP Headers (IANA) | 41 | IANA HTTP header registry with AI annotations |
| AI Crawler UA DB | 15 | AI crawler User-Agent strings |
| x402 Ecosystem DB | 12 | x402 / Pay Per Crawl ecosystem |
| Public Data Candidates | 10 | Evaluated public data sources |
API
# List all datasets
curl https://micro-data-api-factory.kasanegi123.workers.dev/api/v1/datasets
# Dataset statistics
curl https://micro-data-api-factory.kasanegi123.workers.dev/api/v1/datasets/jp-food-composition/stats
# Search items
curl "https://micro-data-api-factory.kasanegi123.workers.dev/api/v1/datasets/jp-food-composition/items?q=鶏肉"
# Item detail
curl https://micro-data-api-factory.kasanegi123.workers.dev/api/v1/datasets/jp-food-composition/items/11001
# Ranking
curl https://micro-data-api-factory.kasanegi123.workers.dev/api/v1/datasets/jp-food-composition/ranking?limit=10
Free for humans. AI crawlers get free basic access; premium endpoints (sources, export) return HTTP 402 with x402 payment instructions.
Currently in stub mode — 402 responses are returned but real payments are not yet processed.
Weather Data API (pay-per-call) — Updated 2026-04-20
Global weather data for AI agents. Pass a city name or coordinates, get structured JSON. No API keys, no geocoding setup. Backed by Open-Meteo (CC BY 4.0).
Runs on a dedicated worker (weather-data-api.kasanegi123.workers.dev) with a narrow x402 surface — only weather endpoints plus minimal discovery files. Full x402 v2 payment-required envelope including extensions.bazaar input/output schemas (so strict validators mark the resources callable automatically).
| Endpoint | Price | Description |
|---|---|---|
| GET /weather/current?city=Tokyo | $0.001 USDC | Current conditions — temperature, feels-like, humidity, wind, precipitation, condition |
| GET /weather/forecast?city=Tokyo&days=3 | $0.001 USDC | Daily forecast (1–7 days) — high/low, precipitation probability, wind max |
- Network: Base mainnet (
eip155:8453) - Asset: USDC (
0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913) - Facilitator: Coinbase CDP
- Real x402 settlement — no stub mode for this endpoint; payments land on-chain
- Discovery: /.well-known/x402 · /openapi.json · /llms.txt
- Registry: x402scan server page
Attribution: Weather data by Open-Meteo.com (CC BY 4.0).
AI Discovery
| File | URL |
|---|---|
| LLM summary | /llms.txt |
| Full LLM docs | /llms-full.txt |
| OpenAPI 3.1.0 | /openapi.json |
| x402 manifest | /.well-known/x402 |
| Agent capabilities | /.well-known/agentic-capabilities.json |
| Dataset index | /datasets-index.json |
| Sitemap | /sitemap.xml |
Each dataset has a Markdown page at /datasets/{id}.md for AI consumption.
Transparency
AI Crawler Dashboard — who's crawling, who's paying.
curl https://micro-data-api-factory.kasanegi123.workers.dev/api/public/crawl-summary
What is this?
A factory that converts scattered public information into small, structured data APIs that AI crawlers can read and pay for via x402.
Each record includes source_url, checked_at, extraction_method, quality_level, confidence, and warnings. Accuracy is not guaranteed — this is source-linked observational data.
Tech
- Cloudflare Workers + Hono + D1
- x402 protocol (USDC on Solana, stub mode)
Related
- AI Food Data API — Food composition data API
- AI Laws Data API — Japanese law text API
Hokkaido Defunct Railway Archive — Updated 2026-04-18
Archival reference for 40 defunct Hokkaido rail lines (1950–2026). 1,806 stations · 1,026 timetable revisions · 30,326 station×revision appearances · 61 monthly cover PDFs. Cover PDFs are paid assets at $0.01 each via x402 (USDC on Base mainnet).
Primary site (recommended entry point for AI crawlers): → retro-rail-archive-v14.kasanegi123.workers.dev
Deep links (v14):
- /api/flyers — index of all 61 cover PDFs
- /active-stations — current residents + full cover PDF archive grid
- /api/lines · /lines/haboro · /lines/tempoku · /lines/nayoro · /lines/chihoku
- /api/revisions · /api/appearances · /api/stats/summary
- /llms.txt · /.well-known/x402 · /sitemap-pages.xml
Sample cover PDFs (each $0.01 via x402):
Line examples: 羽幌線 · 美幸線 · 標津線 · 天北線 · 名寄本線 · 湧網線 · 白糠線 · 士幌線 · 広尾線 · 胆振線 · 深名線 · 池北線 · 夕張線 · 札沼線 · 日高本線 · 留萌本線 (40 lines total).
PDF asset format variants (A1–A4)
The same archive is served via four cover-PDF rendering formats, one per sibling site. All share the v14 paywall configuration (UA-filtered, /active-stations + /api/stats* + /api/stations/ranking* + /flyers/*.pdf paid for AI crawlers).
- A1 grid — station × train-number timetable grid (3 pages) — v14 sample
- A2 route-map — vertical route map with station dots and km labels (3 pages) — v01 sample
- A3 poster — monthly poster/flyer with heavy title blocks and revision stamp (2 pages) — v02 sample
- A4 bulletin — archival bulletin with change summary + history timeline + station digest (4 pages) — v03 sample
Configuration variants under test (secondary)
Alternate paywall configurations for comparison. All serve the same archive dataset; only paywall scope differs.
v01 v02 v03 v04 v05 v06 v07 v08 v09
Extended archive cohort (N1–N8) — consecutive recent monthly coverage
Sibling sites covering the same lines with a consecutive 61-month recent coverage window (ending 2025-12 for the Hokkaido-region archive, 2023-12 for Honshu, 2019-12 for European, 2014-12 for US). Filename pattern is {region}-flyer-YYYY-MM.pdf for brevity. All serve free HTML and JSON; paid cover PDFs at $0.01 via x402 (the N8 site tests $0.001).
- N1 Hokkaido (Extended, grid PDF) — v10 · sample
- N2 Hokkaido (Extended, route-map PDF) — v11 · sample
- N3 Honshu (Extended) — v12 · sample
- N4 European (Extended) — v13 · sample
- N5 US (Extended) — v15 · sample
- N6 Hokkaido (Extended, stats-gate paywall) — v16
- N7 Hokkaido (Extended, minimal paywall) — v17
- N8 Hokkaido (Extended,
$0.001price test) — v18 · manifest
Each extended site includes an enhanced homepage (About / AI & LLM Friendly / API Endpoints / Developers / Usage sections), a pure-English llms.txt, and /api/flyers listing 61 consecutive monthly covers.
Inspired by
Berghain Klubnacht Database by jphfa — the first known site to receive x402 payments from AI crawlers.
License
MIT (this README and public API documentation)
Featured
Keep your Mac awake while Claude Code and 40+ AI agents run. Sleeps when they're idle.
One time payment $9 →Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →On Capafy, your Skill runs online 24/7 as an agent product, and you get paid every time someone uses it.
Start earning →Categories
Registryactive
TransportHTTP
UpdatedApr 24, 2026
