# Changelog of Job Scraper for ZipRecruiter 💰 $0.99/1K — 7 Country Markets (`blackfalcondata/ziprecruiter-job-scraper`) Actor

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

## Changelog

### 0.1.x — 2026-08-13

- Fixed: **a momentary error at the source could end a whole run.** One failed
  request from ZipRecruiter's own backend was enough to stop a run with no results,
  even though a retry seconds later would have worked. Such hiccups are now retried,
  and if the source really is down the run says so instead of reporting an unknown
  error.

### 0.1.x — 2026-08-12 (c)

- **Removed: `companyWebsite` and `companyLogoUrl`.** Both were listed as company
  metadata and both were always empty — 0 of 716 records checked across every market.
  If you read these fields, they were already `null` on every run; they are now
  absent. `company` is unaffected.
- Fixed: **a US search with no location returned nothing.** Leaving Location blank
  means "search the whole country", but on the US market it produced an empty result
  set on an otherwise successful run. It now searches nationwide as intended. Other
  markets were never affected.

### 0.1.x — 2026-08-12 (b)

- Fixed: **`description` held raw HTML, and `descriptionHtml` was always empty** —
  the three description fields were inverted on every market. `description` is now
  plain text, `descriptionHtml` carries the HTML, and `descriptionMarkdown` is real
  Markdown. This is what made spreadsheet/CSV exports look like the columns had
  shifted: a multi-line HTML blob sat in the column you expected prose in.
- Fixed: **`descriptionFormat: "html"` returned records with no description at all**,
  because the field it keeps was always empty. All four modes now return content.
- Fixed: **incremental mode re-reported unchanged US jobs as updated on every run.**
  The apply link carries a per-visit token that changed each time, so change
  detection saw every job as modified — a 200-job run re-delivered 137 unchanged
  records. A repeat run with no real changes now delivers none.

### 0.1.x — 2026-08-12

- Fixed: **job descriptions were missing from most records in larger US runs**,
  even with "Include Full Details" enabled — a 474-job run delivered only 31
  descriptions and the rest came back `null`. Descriptions are now filled across
  the whole result set.
- Fixed: **searches in the non-US markets (IE, GB, DE, FR, IN, NZ) started
  failing** around 2026-08-10 with a "blocked / retry later" error. All seven
  markets return results again.
- Improved: memory headroom — the run no longer risks being killed without an
  error message when a large result set is collected in a small memory bucket.

### 0.1.x — 2026-07-07

- Improved: reliability against intermittent network hiccups — a transient
  connection error early in a run is now retried instead of ending the run
  with no results.

### 0.1.x — 2026-06-08

- Added: full job **descriptions for the US (.com) market**. Previously US listings returned only a short snippet (`descriptionSnippet`); now `description` / `descriptionMarkdown` carry the complete job text, fetched from the detail page. `includeDetails` now applies to US as well as the other markets.
- Improved: US search reliability — transient render failures are now retried instead of failing the whole run.

### 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 for change detection)

### 0.1.x — 2026-04-14

- Added: cross-run repost detection (`isRepost`, `repostOfId`, `repostDetectedAt`)
- Added: `skipReposts` input to exclude detected reposts from output

### \[0.1.0] - 2026-04-05

#### Added

- Initial release with 7 markets (US, GB, IE, IN, DE, FR, NZ)
- US market: salary, benefits, remote status, apply URL via RSC parsing
- IE/GB/IN/DE/FR/NZ: full description + geo coordinates via JSON-LD
- Incremental mode with cross-run change detection
- Compact output mode for AI agent workflows
- Parallel page fetching for fast pagination