Eures Jobs Search Scraper

Searches the official public EURES job API and sequentially verifies each selected detail response. Only complete, identity-matched detail records with readable descriptions and source-quality checks are written to the dataset. Records include source-backed identifiers, title, employer, locations, openings, contract and schedule codes, categories, employment period, remote-work flag, readable HTML/text/paragraph descriptions, requirements, remuneration, application metadata, dates, language profile, public portal URL, and direct-access quality evidence. Request receipts and run diagnostics are stored in the key-value store; job records retain their associated source-provenance receipts. Recruiter contact objects and raw contact payloads are not collected; contact details in descriptions and application instructions are redacted.

The public input contains only query, bounded maxItems and maxPages, optional countryCode, and language. The Actor uses ordinary direct HTTPS with a fixed internal timeout, no proxy, no custom user-agent, no retries, no stealth/fingerprint changes, no alternate identity, and sequential search/detail requests. HTTP 401/403/429/451 or a visible source-authored access challenge stops requests, discards staged jobs, and records platform status: SUCCEEDED with resultStatus: SKIPPED. HTTP 407, unverified redirects, generic transport failures, and unverified source responses are DEFERRED, not proof that EURES blocked access. NO_DATA is emitted only when a successful official search response includes both an empty jvs array and an explicit zero numberRecords; the receipt-backed evidence is stored in OUTPUT_SUMMARY.emptyResultEvidence. An unrecognized/empty response without both facts is DEFERRED. Fatal non-block errors with no stored jobs call Actor.fail() and report FAILED / FAILED. If a later non-block failure follows successful dataset writes, those jobs remain SUCCEEDED / LIMITED. OUTPUT_SUMMARY and OUTPUT separate platform status (SUCCEEDED or FAILED) from extraction resultStatus.

Local development

npm test
npm run check
apify validate-schema

These checks are offline and do not start an Actor or read its local storage. npm run validate:dataset is storage-backed and is intentionally not part of the lifecycle audit's safe check set. Do not run an Actor, purge storage, or inspect generated datasets unless that separate operation is explicitly authorized and uses a deliberately selected storage directory.