Seit 25 Jahren unterstützen wir bei inovex mit rund 500 IT-Expert:innen Unter…", "text": "Seit 25 Jahren unterstützen wir bei inovex mit rund 500 IT-Expert:innen Unterneh…" }, { "name": "description", "title": "Was du bei uns bewegen kannst", "html": "
Agile Projektarbeit: In cross-funktionalen Teams…", "text": "Agile Projektarbeit: In cross-funktionalen Teams setzt du Projekte für Kunden un…" }, { "name": "profile", "title": "Wer gut zu uns passen würde", "html": "
Dann möchten wir dich kennenlernen!
\nDas People &…",
"text": "Dann möchten wir dich kennenlernen!\n\nDas People & Culture Team ist dein Ansprech…"
}
],
"description": "Seit 25 Jahren unterstützen wir bei inovex mit rund 500 IT-Expert:innen Unternehmen bei der Digitalisierung und Agilisie…",
"benefits": [
"Kfz-/Reisezulage",
"Coaching-Angebote",
"Firmenwagen",
"Mitarbeiter-Laptop",
"Betriebsarzt im Unternehmen",
"Firmenhandy",
"Betriebliche Altersvorsorge",
"Mitarbeiterrabatte",
"Zusätzlicher Urlaubstag",
"Fitnessangebote",
"Flexible Arbeitszeiten",
"Gute Verkehrsanbindung",
"Homeoffice",
"Firmenevents- und Veranstaltungen",
"Vorsorgeuntersuchung",
"Parkplatz",
"Sabbatical",
"Sportaktivitäten",
"Studienförderung"
],
"applicationType": "ThirdParty",
"externalApply": true
}
```
### Run summary
Every run leaves an **OUTPUT** record (JSON) and a readable **Run summary** (HTML, Output tab). For each start URL you see the status, the jobs delivered, the distinct jobs seen versus the number the site announced, the pages that failed, and plain-English warnings (`anomalies`).
| Status | Meaning |
|---|---|
| `done` | Every announced page was read and the **main results** seen match the site's announced count (nearby jobs count towards it only when `includeNearbyJobs` is on). On lists of 2+ pages a gap of up to 1.5 % is tolerated, because Stepstone reorders results while you page through them (19 of 2,129 measured); on a single-page list no gap is tolerated. The summary always prints "X of Y announced main results seen" and, separately, the nearby jobs. |
| `partial_collection` | You did **not** get the full listing: a page failed after retries, came back empty, only repeated earlier jobs, the page count shrank during the run, or the gap is larger than normal reordering. Numbers are in the summary and an `ANO_COUVERTURE` warning. |
| `stopped_at_site_limit` | The site announces more jobs than its pages can show. Split the search (smaller radius, more filters, several cities). |
| `stopped_at_limit` | Your `maxItems` or maximum charge was reached. |
| `no_results` | The site announced 0 results. Warning `NO_RESULTS`: the search is empty, the URL is wrong or the site is down. |
| `no_new_jobs` | Every job of this URL was already delivered by an earlier start URL. |
| `job_not_available` | A job page URL returned something else (a search listing, or another job): nothing is delivered or charged. |
| `not_processed` | Not reached: your `maxItems` or maximum charge was reached first. |
| `interrupted` | The run stopped (time limit or abort) while this URL was being read; resurrect the run to continue it. |
| `not_found` | The site answered 404: the search or job page does not exist (a removed job, a mistyped URL). Read once, nothing delivered or charged. |
| `blocked_by_site`, `rate_limited`, `unrecognized_page`, `structure_changed`, `network_error`, `http_error`, `site_maintenance`, `internal_error` | The site refused every proxy session, limited the requests, served something that is not a job listing or a maintenance page, or the Actor hit an unexpected error. This is **never** reported as an empty search. |
If **no** start URL delivers a job, the run **fails** with a message that names the cause. One empty URL next to URLs that deliver only adds a warning. With details on, job pages that could not be read are counted (`detailErrors`, warning `ANO_DETAILS`) and those jobs are delivered without detail fields. If 5 job pages in a row are refused by the site, the Actor stops reading job pages for the rest of the run and delivers the remaining jobs without details (`detailsSkipped`, warning `DETAILS_STOPPED`). The summary never says "all fine" in any of these cases.
### Pricing
Pay per result, plus a start fee:
- **Job** (`apify-default-dataset-item`): $0.0009 per delivered job — $0.90 per 1,000 jobs.
- **Job details** (`job-details`): $0.0005 per delivered job **whose job page was actually read** (`detailsSource: "jobPage"`) — only with `includeDetails`, or for a job page URL you pasted.
- **Actor start** (`apify-actor-start`): $0.005 per run (once per GB of memory; the default is 1 GB).
Apify subscribers pay less automatically, on every event: 10% off on the Bronze plan, 20% on Silver, 30% on Gold and above.
| Run | Price |
|---|---|
| 20 jobs, list only | $0.023 |
| 20 jobs, with details | $0.033 |
| 1,000 jobs, list only | $0.905 |
| 1,000 jobs, with details | $1.405 |
These totals include the $0.005 start fee.
A job delivered without its details (job page refused, removed, or details stopped after 5 refusals in a row) costs only the job price. Searches that return nothing, suggestions and duplicates are not charged. The final message and the Run summary give the number of `job-details` events charged (`detailsCharged`).
**Maximum charge per run**: with details on, the Actor only reads a job page if both the job and its details still fit in your limit. When the limit no longer covers a job **with** its details, the run stops cleanly (`stopped_at_limit`, "maximum charge per run was reached") instead of delivering jobs without the details you asked for; a few cents may remain unspent when some job pages could not be read. If a run is migrated by the platform exactly between writing a batch of jobs and charging its details, those details are not charged (never charged twice).
### Known limits
- **Access**: stepstone.de refuses datacenter proxies; the default is Apify residential proxy, Germany. Some residential IPs are refused too; the Actor switches to a new IP on each refusal (up to 8 per request) and retries random server errors, then gives each page one last try after a pause. A page that still fails is skipped, counted and reported (`partial_collection`), never silently dropped. If you turn the proxy off or choose a non-residential group, the Actor respects it and says so in the final message and the Run summary (`PROXY_SETTING`).
- **Speed**: 3 to 5 seconds between requests. A list page holds 25 jobs, so 1,000 jobs take about 40 pages (~3-4 minutes). **With details, each job costs two more requests** (job page + apply-button API): 500 jobs take about 70 minutes, 1,000 about 2.3 hours. The default run timeout is **2 hours** (enough for the default 500 jobs with details, with margin); raise it, or lower `maxItems`, for bigger runs with details.
- **Completeness**: judged on the main results only. Stepstone reorders results during paging; a full read of 2,129 jobs missed 19 (0.9 %). Beyond 1.5 % (any gap on a single-page list) the URL is reported `partial_collection`.
- **Salary**: always Stepstone's estimate. The employer's own salary, when any, is only inside the description text. A few estimates give only a lower bound (`salaryMax` null).
- **Address**: given by 19 of 31 job pages checked; sometimes only a city, sometimes in English ("80331 Munich, Germany"). Delivered as written.
- **Application link**: `externalApply` says whether you apply on the company site; the external URL itself is behind a Stepstone redirect and is not delivered.
- **Not covered**: logged-in content (salary behind "Gehalt anzeigen"), Stepstone sites of other countries.
### Support
Check the **Run summary** first. For a problem you can reproduce, open an issue in the Actor's **Issues** tab with the run ID, the start URLs and what you expected versus what you got (never paste API tokens).
This Actor is maintained by Hexagone Data. It is not affiliated with, endorsed or sponsored by StepStone GmbH; "StepStone" and "Stepstone" are trademarks of their owner, used here only to describe the website the data comes from. It reads public job pages without logging in; use the data in line with the site's terms and applicable law (personal data in job descriptions, such as contact names, is delivered as published).
# Actor input Schema
## `startUrls` (type: `array`):
stepstone.de search pages, with your filters exactly as the site built them (e.g. https://www.stepstone.de/jobs/python or https://www.stepstone.de/jobs/pflegefachkraft/in-hamburg?radius=50), or single job pages (stellenangebote--...-inline.html). Pagination is followed up to the last page announced by the site. When set, keyword/location/radius are ignored.
## `keyword` (type: `string`):
Job title or skill, e.g. python or pflegefachkraft. Builds https://www.stepstone.de/jobs/