Japan Government Tenders Scraper — official procurement notices from all of Japan

Search and extract Japanese government procurement notices (tenders / bids / RFPs) — national ministries, all 47 prefectures, municipalities and independent agencies — from the official 官公需情報ポータルサイト (Kankōju public procurement portal, run by the SME Agency). Think SAM.gov for Japan: keyword search, prefecture and category filters, publication-date ranges, full notice text and source PDF links, returned as clean structured JSON.

No login, no CAPTCHAs, no Japanese needed to operate it — and the underlying source is an official government API, so the data is legal to use and exceptionally stable.

What data does it extract?

Field	Description
noticeKey	Unique ID of the notice inside the portal
title	Tender title (Japanese)
organization	Procuring organization (e.g. 国立大学法人岡山大学, 東京都小平市)
prefectureCode / prefectureName	JIS prefecture code and name
cityCode / cityName	Municipality code and name
category / categoryJa	goods / construction / services (+ original Japanese)
procedureType	Procedure type (e.g. open competitive bidding) when published
certificationLevels	Required bidder qualification grade (A–D) when published
publishedAt	Publication date (ISO 8601)
tenderStartAt / tenderOpeningAt	Bid submission start / bid opening date when published
deliveryDeadline	Delivery / performance deadline when published
location	Place of performance
description	Full notice text — requirements, qualifications, contacts, deadlines
sourceUrl	Original notice document (PDF/HTML)
attachments	Attached files (specifications, forms) with names and URLs
scrapedAt	ISO timestamp of extraction

Sample output

{
  "noticeKey": "dG9reW8va29kYWlyYV9jaXR5LzIwMjYvMjAyNjA3MDZfMDAwMTcK",
  "title": "上水南保育園給食調理業務委託事業者募集（公募型プロポーザル）",
  "organization": "東京都小平市",
  "prefectureCode": "13",
  "prefectureName": "東京都",
  "cityName": "小平市",
  "category": "services",
  "categoryJa": "役務",
  "publishedAt": "2026-07-06T00:00:00+09:00",
  "description": "上水南保育園給食調理業務委託事業者募集（公募型プロポーザル）…",
  "sourceUrl": "https://www.city.kodaira.tokyo.jp/kurashi/files/128142/128142/att_0000001.pdf",
  "attachments": [
    { "name": "仕様書（案）（PDF 325.3KB）", "url": "https://…att_0000002.pdf" }
  ],
  "scrapedAt": "2026-07-07T00:53:30.785Z"
}

Input

All filters combine with AND. At least one of query / projectName / organizationName / prefectures is required.

• query — full-text search with boolean operators: システム AND 保守, 清掃 OR 警備, 工事 ANDNOT 解体
• projectName / organizationName — partial match on tender title / procuring organization
• prefectures — JIS codes (13), romaji (tokyo) or Japanese (東京都)
• category — goods (物品) / construction (工事) / services (役務)
• publishedFrom / publishedTo — publication-date window (YYYY-MM-DD)
• maxItems — cap on results; you are only charged for what you get

The portal's API caps a single query at 1,000 results — this Actor transparently pages beyond that by slicing the publication-date window, deduplicating as it goes.

Pricing (pay per event)

Usage	Approx. cost
100 tenders	$0.22
1,000 tenders	$2.02
10,000 tenders	$20.02

Actor start: $0.02 + $0.002 per tender. No subscription, no rental fee. A typical daily monitoring run ("yesterday's new notices, one keyword, one prefecture") costs a few cents.

Common use cases

• Bid intelligence & lead generation — monitor new tenders in your industry or prefecture daily, pipe them to Slack/CRM before competitors read the portal
• Market research — quantify Japanese public-sector demand by region, category and buyer (hundreds of new notices per day)
• AI agents / RAG pipelines — full notice text in every item makes this a drop-in source for LLM summarization, matching and alerting

Use from AI agents (MCP)

This Actor can be called as an MCP tool via Apify MCP server. Point your agent to mcp.apify.com and search for this Actor by name.

FAQ

Is this legal? Yes. The data comes from the Japanese government's official public API (operated by the SME Agency), published precisely so procurement information can be reused programmatically. No personal data, nothing behind logins.

Why are some fields null? The portal aggregates notices from hundreds of source systems; procedureType, certificationLevels and some dates are populated only when the source provides them. title, organization, publishedAt, description and sourceUrl are reliably present.

Results are in Japanese — can I use this without speaking Japanese? Yes. Filter with romaji prefecture names and English category values, then pipe title/description through any translation or LLM step.

How fresh is the data? The portal ingests notices continuously; each run queries the live API at execution time.

Something broken? Open an issue on this Actor's page — fixes ship within 24 hours.

---

Maintenance notes (internal): source = https://www.kkj.go.jp/api/ (GET, XML, token-free), spec = https://www.kkj.go.jp/doc/ja/api_guide.pdf (V1.1, verified 2026-07-07). Required: one of Query/Project_Name/Organization_Name/LG_Code, else <Error>no searchword</Error> (HTTP 200). Legitimate empty = SearchHits=0. Max 1,000/call, no pagination param → CFT_Issue_Date bisection. No auth, no rate limiting observed; self-throttle 300 ms + exponential backoff.