# Changelog of Arbeitsagentur Job Scraper — Salary, Contact & Details (`haketa/arbeitsagentur-scraper`) Actor

- **URL**: https://apify.com/haketa/arbeitsagentur-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/haketa/arbeitsagentur-scraper.md

## Changelog

### 1.0.1 — 2026-11-01

- **Fresh verification run**: scraped **500 unique jobs (nationwide `Pflegefachkraft`, full details) in ~20 s with 0 duplicates**. Core/identity fields (`referenceNumber`, `title`, `occupation`, `offerType`, `employer`, `city`, `region`, `country`, `latitude`, `longitude`, `portalUrl`, `description`, `datePosted`, `startDate`) at ~100% coverage; pagination across the v6 API confirmed working.
- **Bug fix — `allianzPartnerUrl` now a usable link**: the detail API often returns partner URLs as bare domains (e.g. `helixjobs.com`, `www.drk-hannover.de`). These are now normalized to absolute `https://` URLs (already-schemed URLs are left untouched), so every `allianzPartnerUrl` value is a valid, clickable link. Fixed 74/100 scheme-less values in the verification sample.
- **README refreshed to the gold standard**: output example and full field-reference table regrounded in the fresh sample (real reference numbers, `JAHRESGEHALT`/`STUNDENLOHN` salary formats, normalized partner URLs); added an honest data-dependent coverage note; corrected the input table to match the real schema (added the `city` dropdown, removed a non-existent `language` parameter); expanded the Pagination & scale and FAQ sections.
- Re-confirmed all v6 search + v4 detail endpoints, salary period mapping, contact/requirements extraction and multi-location output against the live Bundesagentur API. No input/output field names changed.

### 1.0.0 — 2026-10-25

- Baseline release. Scrapes arbeitsagentur.de via the public v6 search API + v4 detail enrichment: salary amount/period/type, employer contact mining (email/phone/URL), multi-location array, work-time & contract flags, occupation codes, education & language requirements, and full descriptions. No proxy, browser or login required.
