🔥 Welcome To The Jungle Jobs Scraper
Pricing
from $0.20 / 1,000 job listing with salary & companies
ℹ️ Retrieve jobs, articles, & organizations from "Welcome to the Jungle" using this Actor. Precision meets ease in this modern data tool. Perfect for recruiters, jobseekers, & researchers. Your key to the latest job market insights.
Jobs, articles and organizations from Welcome to the Jungle. Dates are the day each version was released.
Pay-per-event pricing: you pay for the rows you get, and platform usage is included. Apify is retiring rental pricing, so the $20/month subscription is replaced by a price per result — lower on paid Apify plans.
| Price per 1 000 | Free plan | Starter | Scale | Business and above |
|---|---|---|---|---|
| Job listing — every job row | $0.30 | $0.20 | $0.20 | $0.20 |
| Full job details & apply link — add-on, per job | $0.50 | $0.40 | $0.40 | $0.30 |
| Enriched company profile — every company row | $0.40 | $0.30 | $0.30 | $0.20 |
| Article — every article row | $0.40 | $0.30 | $0.30 | $0.20 |
Each run also carries one Actor start event, at $0.00005.
Saved inputs keep working unchanged, and Fetch full job details stays on by default. If you only need listings, switching it off now lowers the bill as well as the run time.
Toolchain update: the Actor now runs on Node 24. It was still built on apify/actor-node:18 — a
Node release that went end-of-life in April 2025 and stopped receiving security patches.
apify/actor-node:24, Apify's current default image), and the Dockerfile was
rebuilt from Apify's current template: correct file ownership, and the process is now started
directly rather than through an npm wrapper, so the platform's stop signal reaches it.The Markdown converter's major version slightly changes how it formats 8 fields in 1 000. Both
changes are corrections: a blank line now precedes a --- horizontal rule (without it, Markdown
parsers read the line above as a heading), and nested list items are indented by two spaces instead
of three, which is the standard. Plain-text output is unchanged, byte for byte — verified across
501 real job descriptions.
Start URLs is now one thing only: paste a page URL, get that page back as a row. It used to do three unrelated jobs at once, which made it hard to predict what a run would return.
/companies/{company}) now returns that company's profile as a row — the
same kind of row the Organizations action produces, and a richer one. It carries all 25 fields of a
native company row plus 10 more: the gender-equality indexes (equality_indexes, present for
two thirds of companies), the company's official website, its canonical and per-language WTTJ
URLs, when the profile was last updated, its cover image and its default language.financo is now Arkéa Financements & Services, solem is Primal —
and the row carries the company's current slug and URL, with requestedSlug recording the one you
pasted so rows still match your input./jobs/anything in particular used to be read as an empty search and
did exactly that. Pasting the /jobs landing page still means "everything", and an unrecognised URL
alongside a real filter still just warns.A company URL used to filter the search to that company, so a run returned that company's jobs. It now returns the company instead. To get a company's jobs, put it in the Companies filter, which takes as many companies as you like and now accepts a full company URL as well as a slug. The run log says so explicitly whenever a company URL is resolved.
The Organizations action used to return a lighter company record, while a pasted company URL returned a fuller one — the same company came back with 25 fields one way and 37 the other. The action now fetches the same company record, so a company row is identical whichever route produced it (verified byte for byte). That adds the gender-equality indexes, the official website, the canonical URLs and the last-updated date to every company the action returns.
It adds very little time — 1 000 companies take about 8 seconds. Set Include details to false
for the previous, lighter row, which is about twice as fast. When a company's profile has been
withdrawn (0.6 % of a 1 000-company run) the row keeps its basic record and says so in detailsError.
The dataset views were rebuilt from scratch. Ten overlapping views became three, one per kind of row, and each one now carries every field that row type has.
Overview (198 columns) holds every field of a job, Articles (28) every field of an article,
Organizations (56) every field of a company. Nothing is split across views any more, and nothing
is missing from them: previously eight of the ten views were different partial takes on a job, and
picking the right one to export from was guesswork.profession.id, which WTTJ does
not serve, and a company cover image under a language level that job rows do not have. All 282
columns are now verified against the live platform with real rows before release, so a column that
the dataset would quietly drop cannot ship again.No change to the data itself: the rows the actor pushes are identical, only the way the dataset presents them changed. Existing exports by field name keep working.
action: "get-jobs") on every run, a five-URL request could not be told apart from a
search and went on to enrich all ~86 000 listings. Pasted URLs are now a complete request on their
own unless a query, filter, date window, scope URL or non-default action is also set.Versions before 2.0 are recorded in the git history.