en-japan Jobs Scraper (Japan)

Scrape job listings from en-japan (エン転職), one of Japan's major mid-career job boards, and get one flat row per job: id and URL, title, company, location, salary as printed (plus a yen band when the text states one), employment type, badges and the job text on the card. Japanese text is kept exactly as en-japan prints it.

Run it for an area (Tokyo is 23), an occupation, a keyword, the day's new arrivals, or paste any search you built on the site.

Example input

{
    "areaIds": ["23"],
    "newOnly": true,
    "maxItems": 500
}

That is the daily feed: the Tokyo jobs en-japan flags as new arrivals (526 of Tokyo's 6,266 on 2026-10-23). To read a whole area past the ceiling below, list occupations too:

{
    "areaIds": ["23"],
    "occupationIds": ["100000", "150000"],
    "maxItems": 4000
}

Input

field	meaning
searchUrls	Optional. en-japan search result URLs on https://employment.en-japan.com/search/search_list/, copied from your browser after setting conditions on the site. Every condition is kept; the actor manages the page number. When given, the four filter fields below are ignored
areaIds	Optional. en-japan area ids (Tokyo is 23). Each area is its own search
occupationIds	Optional. en-japan occupation ids: 100000 is the whole sales group (営業系), 101000 corporate sales (法人営業), 150000 planning, clerical, marketing and admin. Each occupation is its own search, combined with each area
keyword	Optional. A keyword searched across the job text
newOnly	Only jobs en-japan flags as new arrivals (its 新着のみ filter)
fetchDetails	Open every job's detail page to fill posted_from, posted_until and last_updated. One extra request per job; off by default
maxItems	Stop after this many jobs in total, shared across all searches. en-japan shows 50 jobs per page
proxyConfiguration	Apify datacenter proxy by default - see Proxy below

At least one of searchUrls, areaIds, occupationIds, keyword or newOnly is required: en-japan answers a search with no condition with an empty page, so the actor refuses it rather than report "zero jobs". An id the site does not know is ignored by the site, not rejected; check the first rows.

Output

field	meaning
job_id	en-japan's numeric job id, from the /desc_<id>/ part of the job URL (a string)
url	The job's detail page on employment.en-japan.com
title	The job headline
company	Company name as printed
employment_types	The employment-type badges, e.g. ["正社員"]; a card can carry two or three
tags	The other badges: 職種未経験OK, 転勤なし, 完全週休2日, エン エージェント求人 ...
catch_copy	The headline sentence above the card's photo
teaser	The opening of the job text beside the photo (en-japan cuts it with "……")
job_summary	The card's 仕事内容 (work) row
requirements	The card's 応募資格 (who it is for) row
salary_text	The card's 給与 (salary) row as printed
salary_min_jpy, salary_max_jpy	Yen, only when salary_text states exactly one band or floor - see below
salary_period	year for a 年収 / 年俸 band, month for a 月給 band, null when there is no band
location	The card's 勤務地 (work location) row
company_feature	The card's 会社の特徴 row, printed only on エン エージェント cards
editor_comment	The reporter's comment in the card's 取材 box
image_url	The card's first photo
posted_from, posted_until, last_updated	ISO dates from the detail page; only with fetchDetails
search_url	Which search produced this row (page number stripped)

Long rows are cut by en-japan on the search page with "…" or "……"; this actor returns them as printed. Turn on fetchDetails only for the three date columns - it does not add the full job text.

Which fields can be null, and why null is not zero

Counts are from the 2,005 cards on the 40 pages of Tokyo results read on 2026-10-23.

• salary_min_jpy / salary_max_jpy / salary_period are null on 907 of 2,005 cards. A value is filled only when the salary text states exactly one band or floor next to a period word, such as "月給28万円～40万円" or "年収500万円以上". Two bands ("東京：月給21万円～／大阪：月給19.7万円～"), example incomes ("月収例25.6万円") and a bare "350万円～550万円" with no period word are not one band for one period, so they stay null. A floor has no maximum, so salary_max_jpy is also null on those. A 0 would read as "unpaid" and an invented band as a measurement.
• posted_from / posted_until / last_updated are null unless you set fetchDetails, because the search results print no date of any kind. They are also null for a job whose detail page could not be read (the row is still returned, and the run log says which job).
• catch_copy (367 null) and teaser (436 null): cards that print none. company_feature is null on every card that is not an エン エージェント one. image_url is null on 518 cards with no photo.
• job_summary, requirements, salary_text, location are null when the card has no row with that label; every measured card had them except one without a salary row.

The 2,000-job ceiling

en-japan serves at most 40 pages of 50 (2,000 jobs) for any single search, however many it reports. Tokyo says 6,266 jobs; the pages after the 40th come back empty. The actor reads each search until it ends and, when a search stops at the ceiling, says so in the run log and the run's status message. To read past it, split the search: give several occupationIds (each is its own search of up to 2,000), several areaIds, or use newOnly for a daily feed that stays well under it. A job found by two of your searches is returned and charged once.

The first page of a search also shows about five sponsored cards whose jobs reappear later in the results; they are returned once, not twice.

Proxy

The default is Apify's datacenter proxy: on 2026-10-23 a probe from the platform got a full results page from 10 of 10 datacenter exits (residential 9 of 10, Japanese residential 9 of 10). Residential exits are not needed and cost more.

Billing

Pay per event: $0.0021 per job (job-scraped), plus a small actor-start fee ($0.002) per run. Each job is charged once, after it has been written to the dataset. The same job id is never charged twice in one run, even when two of your searches both return it. fetchDetails does not change the price per job.