# Changelog of Career Site Jobs Scraper — Greenhouse, Lever, Workday (`adderleydata/career-site-jobs-scraper`) Actor

- **URL**: https://apify.com/adderleydata/career-site-jobs-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/adderleydata/career-site-jobs-scraper.md

## Changelog

### 0.2 — 25 Sep 2026

- Five more sources — Workable, BambooHR, Personio, JazzHR and Pinpoint — for twelve in all. Each board is read by the same adapter as Adderley Data's Actor for its source, so rows are identical to that Actor's. Board links on all five are recognised, and so are pins such as `workable:acme`; a link on one of their domains that names no board (a Workable job's short link, a company's Personio login address, a vendor's own website) is refused with a message saying what to give instead.
- A company name or domain is now asked of nine ATSs: BambooHR, Personio and JazzHR as well as Greenhouse, Lever, Ashby, Recruitee, Breezy HR and Rippling. That is up to eighteen requests a name, none charged. The three new ones answer a name they do not have with a redirect to their own website, and that is read as "not found": never retried, never charged, never a failed page or a block. Personio's website can answer that redirect with HTTP 429; the answer is read from the response itself, before the page, so it is still "not found".
- Workable and Pinpoint boards are read from a link or a pin only, like Workday's. Workable's servers refuse an address that sends more than a handful of requests in quick succession, and some Pinpoint sites named after well-known brands are demonstrations listing Pinpoint's own made-up sample jobs.
- When a name is looked up, a Recruitee board holding nothing but Recruitee's own sample postings is treated as not found: Recruitee's trial accounts under large brands' names hold nothing else. Given as a link, such a board is an empty board.
- BambooHR, JazzHR and Pinpoint rows get each saved posting's own record or page for the posted date and other facts their board lists leave out, as on those Actors; Workable and Personio boards carry everything, text included. Personio boards are read in English.
- The publication-date filter applies on Workable and Personio. BambooHR, JazzHR and Pinpoint boards list no dates, so it is not applied there, as on Rippling.
- Prices are unchanged.

### 0.1 — first version

- One list of companies, in any mix of forms: a job-board link on Greenhouse, Lever, Ashby, Workday, Recruitee, Breezy HR or Rippling; a source and a board name (`lever:acme`); a company domain or careers-page link; or a company name. Up to 500 entries per run, of which up to 200 domains or names.
- A domain or a name becomes one or two board names (legal words dropped, joined and hyphenated) and is asked of Greenhouse, Lever, Ashby, Recruitee, Breezy HR and Rippling. A source with no board by that name answers "not found": an answer, never retried, never charged, never counted as a failed page or a block. A company found nowhere is named once in the log and in the run summary, which lists what every entry was found as.
- Workday boards are read from their career-site link only, because a Workday board cannot be found from a name.
- Each board is read by the same adapter as Adderley Data's Actor for its source, so rows are identical to that Actor's: the `job.v1` schema, `source` naming the ATS, the source's own ids, dates, places and pay. Workday and Rippling rows get their posting's own record, as on those Actors; the other sources' boards carry everything in the list.
- Filters on title keywords, location, department and publication date, applied on each source as that source can apply them: Workday runs a keyword as a search on the board and states ages only up to 30 days; Rippling lists no dates.
- Incremental mode across sources: `changeType` of NEW, UPDATED, REAPPEARED or EXPIRED, judged by each source's own rules; unchanged postings skipped and not charged. A source answering "no board by that name" does not stop a run from being complete.
- Optional full descriptions with email addresses, phone numbers and profile links redacted by default, and link addresses and images dropped from the HTML.
- The run's output is declared for Console's Output tab, the Run API and AI agents: postings in the default dataset, the run summary in the default key-value store.
- No recruiter names, emails or phone numbers in any field, from any source.