Stepstone Jobs Scraper

Turn StepStone.de searches into clean, structured job data. Give it keywords and a location — or paste full search URLs — and get back a consistent, analysis-ready dataset: title, company, location, salary when the posting shows one, employment type, remote status, posting date, and the full description. Built for recruiting analytics, labor-market research, and job-aggregation pipelines that need dependable output day after day.

Runs on a monitored chassis: an automated canary exercises this actor daily, and failures are diagnosed and patched under human review. Uptime and mean-time-to-repair are published in the badge above.

Try it in 20 seconds

Press Start with the prefilled input (data engineer / Berlin, 20 jobs) — a first run finishes in well under a minute and costs about $0.13. Then raise maxItems or paste your own StepStone search URLs.

What one result looks like

{
  "board": "stepstone_de",
  "board_job_id": "11713465",
  "url": "https://www.stepstone.de/stellenangebote--Data-Engineer-Machine-Learning-Engineer-...-11713465-inline.html",
  "title": "Data Engineer* / Machine Learning Engineer*",
  "company_name": "inovex GmbH",
  "company_url": "https://www.stepstone.de/cmp/de/inovex-gmbh-24494/jobs",
  "location_raw": "Karlsruhe, Köln, München, Hamburg, Stuttgart, Berlin, Erlangen",
  "country_code": "DE",
  "remote_type": "hybrid",
  "employment_type": "full_time",
  "salary_raw": null,
  "posted_at": "2026-08-10T02:01:19.477Z",
  "description_text": "Seit 25 Jahren unterstützen wir bei inovex mit rund 500 IT-Expert:innen …",
  "tags": ["IT, IT-Datenbanken"],
  "scraped_at": "2026-08-15T23:02:56+00:00"
}

Real row from a monitoring run — nothing hand-edited except the trimmed description.

Why this one

• Measured, not promised. Before pricing, we ran 660 results across two time windows on three proxy tiers and shipped the configuration that succeeded on every run. A canary re-runs both input paths every day.
• Same schema across boards. Output is jobs_v1, the field contract shared by all our job-board actors — combine boards without remapping.
• Removed listings handled. Expired postings are detected three ways (structured-data absence, canonical drift, "nicht mehr verfügbar" text) and skipped, never emitted as ghost jobs.
• Job data only. No recruiter or contact-person fields, by design.

Output schema

This actor emits jobs_v1 — the same field contract as every job-board actor we publish, so multi-board pipelines never remap fields.

Field	Type	Description
board	string	Always stepstone_de
board_job_id	string	StepStone's job ID — dedup key together with board
url	string	Canonical posting URL
title	string	Job title
company_name	string	Hiring company
company_url	string/null	Company profile link
location_raw	string	Location as displayed
city, region	string/null	Parsed location parts
country_code	string	ISO country (DE)
remote_type	string/null	onsite / hybrid / remote
employment_type	string/null	full_time / part_time / contract / internship / apprenticeship
salary_min/max/currency/period	number/string/null	Parsed salary
salary_raw	string/null	Salary text as displayed
posted_at	string/null	Posting date (ISO-8601)
description_text	string	Plain-text description
description_html	string/null	HTML description (opt-in)
tags	array	StepStone categories
apply_url	string/null	Direct apply link
scraped_at	string	Scrape timestamp

No recruiter or contact-person data is collected — by design.

Pricing

$0.004 per job result + $0.05 per run start. No hidden compute or proxy line items — results are fetched through reliable German residential proxies included in the price. A 100-job run costs $0.45 total.

Why this price: in measured head-to-head economics, comparable actors charge similar per-result rates but fail roughly a quarter of their runs — runs you still wait on. This actor's proxy configuration was chosen by measurement (two test windows, 660 results): 100% run success. You pay slightly more per listed unit and less per job actually delivered.

Migrating from another StepStone actor?

searchUrls, maxItems, and proxyConfiguration work here the way you already use them — existing run inputs carry over unchanged. keywords + location + radiusKm are available when you'd rather not build URLs.

Changelog

• 2026-08-09 — v0.1 chassis scaffold (pre-release)