Suumo.jp Scraper

Scrape Suumo (suumo.jp), Japan's largest property portal, for rental listings: rent, management fees, deposits and key money, layout, floor area, floor, building age, address, station access with walking times, photos and the listing agency's name. Monitor price changes with recurring updates.

Why this actor

• Search by area: pick a prefecture and optionally JIS city codes; the actor walks suumo's rental result pages forward for every city you list.
• Links mode: paste rental search links (/jj/chintai/ichiran/...) or single room links (/chintai/jnc_.../). Pagination starts at the page named in the link and walks forward.
• Server-side layout filter: room layouts are applied by suumo's own search grammar, so you never download rooms the filter removes.
• Room detail reading: the search page carries the headline fields; a room's own page adds the photo set, the full spec table, the agency's pitch and the agency name.
• Recurring updates: label every room NEW, UPDATED, UNCHANGED, REAPPEARED or EXPIRED against the previous run of the same setup, so a scheduled run only returns what changed.
• Resume: an interrupted run can continue from where it stopped by run id, skipping everything already collected.

Output

Every record is one rental room. Headline fields:

Field	Meaning
id	suumo room id (from the /chintai/jnc_<id>/ link)
rentYen / rentDisplay	monthly rent in yen, and the site's own display form
mgmtFeeYen	monthly management / common-service fee in yen
depositYen / keyMoneyYen	deposit (shikikin) and key money (reikin) in yen
layout / areaSqm	room layout (1K, 1LDK, ...) and floor area in square meters
floor	the room's floor
buildingName / buildingType	building name and type (rental mansion, apartment, house)
address	building address
stations	station access lines with walking times, e.g. Toei Shinjuku Line/Ogawamachi Station walk 1 min (original wording in Japanese)
buildingAge / buildingFloors	building age and total floors
url / imageUrl	room page link and listing photo

Detail fields (when reading room pages is on, or for pasted room links): photos, photoCount, transport, facing, structure, builtMonth, moveIn, dealType, guaranteeCompany, otherInitialCost, otherMonthlyCost, remarks, point, pointText, agencyName and the rest of suumo's spec table.

Sample rows (illustrative, not from a live listing):

id	buildingName	rentYen	layout	areaSqm	address
000000000001	Sample Residence Chiyoda	161000	1DK	26.0	Tokyo, Chiyoda-ku
000000000002	Example Heights Shinjuku	128000	1K	20.1	Tokyo, Shinjuku-ku

Usage

1. Pick the mode.
2. Search mode: choose the prefecture, optionally list JIS city codes (5-digit codes, exactly as suumo uses them in its own links: 13101 Chiyoda-ku, 13102 Chuo-ku, ...). An empty city list walks the whole prefecture. Narrow with the layout codes if needed.
3. Links mode: paste one or more search or room links.
4. Optional: turn on room detail reading for photos, the spec table and the agency name (a per-room surcharge applies). Pasted room links are their own page, so they are always read.
5. Set max rooms to cap the run (the default run caps at the site's own result order; 0 walks until the site runs out).

Notes and limits

• This actor covers the rental side of suumo (chintai). The sale, land and new-build verticals use a different page structure and are not supported; pasting those links fails with a clear message.
• One run covers one prefecture. City codes must belong to the selected prefecture; the run fails fast when one does not.
• Room availability changes hourly on suumo. A room that disappears between runs is reported EXPIRED by recurring updates only after a complete scan.
• The layout filter is the filter suumo itself verifies in its search URL; the actor ships only filters measured against the live site, so no filter silently does nothing.

Send results into your apps (MCP connectors)

Connect an MCP Connector to this actor and results are piped into your apps after each run. The optional input fields mcpConnectors, notionParentPageUrl and maxNotifyListings control delivery; they are never required.