JobScore Jobs Search Scraper

This Actor extracts public JobScore career-board jobs from Atom feeds and their public detail pages. It does not log in, submit applications, use cookies, or access private applicant data.

Modes

Search a public board/feed and enrich the bounded result set:

{
  "mode": "search",
  "feedUrls": [{"url": "https://hire.jobscore.com/jobs/jobscore/feed.atom"}],
  "maxItems": 5,
  "requestDelayMs": 250
}

Detail enrichment is mandatory. Use mode: "single" with one jobUrl or one-item jobUrls, or mode: "multiple" with several jobUrls. careerUrls, companySlugs, and direct URL aliases remain supported. A text query filters title, company, department, location, employment type, description, and structured sections.

Output

Job records include stable IDs, canonical/source URLs, company website/profile/logo, department, location, employment type, Atom dates/categories, plain text and HTML description, headed responsibilities/requirements/benefits, hiring-team profiles, image/video URLs, application URL/provider, record status, and field-quality metadata. Request receipts and run diagnostics are stored only in KVS. Empty optional fields are omitted. Application URLs are reported only; they are never followed or submitted.

Local development

npm ci
npm test
npm run lint
apify validate-schema
$store = 'storage-jobscore-validation'
if (Test-Path $store) { throw "Refusing to reuse existing local store: $store" }
$env:APIFY_LOCAL_STORAGE_DIR = $store
apify run --resurrect --input-file INPUT.json
npm run validate -- $store\datasets\default

The checked-in INPUT.json, INPUT-single.json, and INPUT-multiple.json files provide bounded local smoke scenarios. Each validation run uses a fresh relative local-store directory; existing Actor data is not purged or reused. Only ordinary public responses are used.

Cloud and quality contract

The same bounded contract runs locally and on Apify Cloud. The Actor follows each candidate to its public JobScore detail page and emits only complete jobs with stable identity, company, location, a substantial description, and more than 20 populated meaningful source fields. “About Us” and “The Opportunity” are exposed as separate company/role overviews, and “Nice to Have” content is kept separate from benefits. Platform status is SUCCEEDED or FAILED; resultStatus independently reports COMPLETE, LIMITED, SKIPPED, DEFERRED, NO_DATA, or FAILED. HTTP 401/403/429/451 or a visible source-authored denial/challenge stops requests and discards buffered jobs as SUCCEEDED/SKIPPED; transport failures and unresolved redirects are deferred without being described as source blocks. A fatal zero-row failure calls Actor.fail() as FAILED/FAILED; a later non-block error after stored rows preserves them as SUCCEEDED/LIMITED. Requests are sequential and direct HTTPS with no proxy, cookies, user-agent override, or unbounded redirect following. A single redirect is followed only when its destination remains on an explicitly allowed public JobScore feed/job route; each hop is retained as its own single-attempt KVS receipt. Fetch or parse diagnostics never mix with job rows.