# Changelog of StepStone.de Job Scraper 💰 $1.50/1K — Salary & Skills (`blackfalcondata/stepstone-de-scraper`) Actor

- **URL**: https://apify.com/blackfalcondata/stepstone-de-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/blackfalcondata/stepstone-de-scraper.md

## Changelog

All notable changes to this actor.

### 0.1.x — 2026-09-11

- Fixed: Posted Within now holds for every delivered job. A job posted outside the window, or without a posting date, is never delivered.
- Fixed: a search never delivers more jobs than StepStone.de reports for it.
- Changed: Minimum Salary was silently ignored. StepStone.de states a salary on too few listings to filter on it, so a run with it set now stops with a message instead of returning jobs that may not meet it.
- Changed: the salary sort options were removed. StepStone.de does not sort by salary, and no run used them.
- Fixed: when a search has only a few matches, StepStone.de fills the page with recommended jobs that ignore your search term and Posted Within. Those are no longer delivered or charged.
- Fixed: a search the site fails to answer now ends the run with an error instead of reporting "no results".
- Fixed: the Contract Type filter returned no results at all. Every contract type is now sent in the form StepStone.de expects (Full Time and Part Time use its working-hours filter). Mini Job and Volunteer have no StepStone.de equivalent, so a run that asks for them stops with a message instead of delivering nothing.
- Fixed: Posted Within, Experience Level, Work Type, Language and Sort by Date are now actually applied; before this, Posted Within, Language and Sort by Date were ignored and Experience Level and Work Type returned no results. A Posted Within value other than 1, 3, 7 or 14 days is honoured exactly.

### 0.1.x — 2026-04-24

- Fixed: detail-worker browser sessions are now isolated, so a single worker refresh no longer replaces the session used by the rest of the pool.
- Fixed: consent handling now waits briefly for cookie state to settle before the session is reused.
- Fixed: pricing and memory metadata are now published from one canonical configuration.

### 0.1.x — 2026-04-14

- Added: `descriptionHtml`, `descriptionMarkdown` output fields (triple-format descriptions for RAG/LLM pipelines)
- Added: `contentHash` output field (SHA-256 hash of content-identifying fields)

### 0.1.21 — 2026-04-14

- Non-blocking worker pool: first worker starts fetching immediately; additional workers join the queue as they become ready.
- Pool size kept at 3 after memory profiling confirmed comfortable headroom at default allocation.

### 0.1.20 — 2026-04-14

- Faster session initialisation with parallelised setup rounds and stagger/backoff to stay under rate thresholds.
- Same-source pairing between initial session and page fetches for improved reliability.

### 0.1.19 — 2026-04-14

- Broader session-readiness detection to cover newer challenge variants.
- Session initialisation now simulates light page activity before readiness polling.
- Readiness poll window extended from 5 s to 15 s.

### 0.1.18 — 2026-04-13

- Tuned initial page navigation strategy to avoid long-tail analytics hangs. Initial load timeout set to 20 s.

### 0.1.17 — 2026-04-13

- Session pool pattern: one ready session per worker, refresh on 403.
- 20-minute session refresh interval to keep long runs healthy.

### 0.1.16 — 2026-04-13

- Parallel search-result pagination (page 1 sequential for `totalCount`, later pages parallel) — up to 5× speedup.

### 0.1.15 — 2026-04-13

- Incremental mode: emit NEW / UPDATED / UNCHANGED / EXPIRED records keyed by `stateKey`.
- Repost detection across runs via expired-entry matching.

### 0.1.14 — 2026-04-12

- Initial session-continuity build with structured data extraction.