SUUMO Japan Rental Listings Scraper

Rental listings from SUUMO, Japan's largest property portal — with every yen figure parsed into a real number.

SUUMO writes prices the Japanese way: rent as 16.4万円, fees as 15000円, and an undisclosed cost as -. This actor returns 164000, 15000 and null, so you can sort and filter without touching the data. It also adds the two numbers people actually compare on:

• monthlyTotalYen — rent plus management fee, what the tenant really pays each month
• rentPerSqmYen — yen per square metre, the honest way to compare a 1K against a 2LDK

One row per available room, with its building's details denormalised onto it — so a single CSV needs no joining.

HTTP-only. No browser, no login.

---

Why it is cheap to run

SUUMO's listing pages are two-level: one card is a building, and inside it a table lists every available room. A single request in testing returned 30 buildings and 156 rooms.

Most job and directory scrapers cost one request per row. This one costs about one request per 150 rows, so a 1,000-room extract is roughly 7 requests.

---

What you get

Field	What it is
roomUrl, roomId	The specific room's SUUMO page and its id
buildingName, buildingType, address	e.g. アーバネックス千代田淡路町 / 賃貸マンション / 東京都千代田区神田小川町１
rentYen, rentRaw	160000, and the original 16万円
adminFeeYen, monthlyTotalYen	Management fee, and rent + fee combined
depositYen, gratuityYen	Deposit (敷金) and key money (礼金) — null when not charged, never 0
layout, areaSqm, floor	1DK, 25.13, 4
rentPerSqmYen	Derived — rent divided by area
buildingAgeYears, buildingFloors	築5年 → 5; 新築 (newly built) → 0
accessLine1..3, accessStation1..3, accessWalkMin1..3	Up to three lines, split into line, station and walk minutes
imageUrl, sourceUrl, scrapedAt, fieldsPopulated	Provenance

Measured fill rates

From a live 156-room page in Chiyoda, Tokyo. Fee disclosure varies a lot by region — a live Osaka run returned deposits on only 23% of rooms against Tokyo's 71%, because the two markets use 敷金 / 保証金 differently. Rent, layout, area and station walk are 100% everywhere tested:

Field	Fill
roomUrl, buildingName, address, rentYen, layout, areaSqm, floor	100%
buildingAgeYears, accessStation1, accessWalkMin1	100%
gratuityYen	84%
adminFeeYen	79%
depositYen	71%
accessStation3	96%

A fee the listing does not charge stays null rather than becoming 0 — a 0 deposit and an undisclosed deposit are very different things when you are budgeting.

---

Input

{
  "prefecture": "tokyo",
  "maxItems": 150,
  "maxRentYen": 150000,
  "maxWalkMinutes": 10
}

Or paste a SUUMO search URL with filters you already set on the site:

{
  "startUrls": [{ "url": "https://suumo.jp/jj/chintai/ichiran/FR301FC001/?ar=030&bs=040&ta=13&sc=13101" }],
  "maxItems": 500
}

Defaults: maxItems is 100 — a single page that finishes in seconds. Raise it for a real extract; one ward of Tokyo alone advertises over 32,000 listings.

Filters

maxRentYen, minAreaSqm, layoutContains, maxWalkMinutes and maxBuildingAgeYears all run before a row is charged, so you are never billed for rows a filter excluded.

Monitoring

Set incremental: true with its own stateKey to emit only rooms not seen before, tagged changeType: "new" — a daily watch on a saved search bills for what actually appeared.

---

Limitations — read before you buy

• SUUMO rate-limits by IP. It answers 503 with a short body when throttling. The actor detects that specifically, backs off and retries, and reports throttleResponses in the run summary. Keep Apify Proxy enabled — it is on by default and included on every Apify plan. A single unproxied address starts drawing 503s quickly.
• A prefecture is not a search. SUUMO addresses a ward by slug (/chintai/osaka/sc_osakashikita/), and a prefecture-wide URL returns HTTP 200, the right heading and zero listings — it is a search-refinement form, not results. Choosing a prefecture therefore runs its ward index and expands every ward (67 for Osaka), so a whole-prefecture run is many searches rather than one. Paste a search URL when you want exactly one area.
• Bus routes have no station walk time. An entry like バス10分 停歩5分 means 10 minutes by bus then 5 on foot from the bus stop. Reporting that 5 as the station walk would make a far-out property look commutable, so accessWalkMin is null for bus access — and maxWalkMinutes therefore excludes those rooms.
• Rentals only. This actor covers 賃貸 (rental). Sale listings use different pages and are not included.
• Japanese text is returned as-is. Building names, addresses, layouts and station names come back in Japanese, because that is what the source publishes and transliterating would introduce errors.
• Duplicate buildings. SUUMO itself notes that the advertised total counts agency postings, and it merges listings it believes are the same property. Rows are deduplicated on room id within a run.

---

Output

Three ready-made views on the Output tab: Overview, Move-in costs (every yen figure, for rent benchmarking) and Stations & commute (all three lines with walk times). A RUN_SUMMARY record reports rows pushed, rows billed, rows filtered out and how often SUUMO throttled.