FreeJobAlert Jobs Search Scraper

Extract current public government recruitment articles from FreeJobAlert. The actor discovers article links across bounded search/update pages, enriches each article from public JobPosting/NewsArticle JSON-LD and page tables, and writes one complete record per notification.

Input

query filters recruitment titles; use latest for all current updates. maxItems is capped at 500 and maxPages at 20. Timeout, request budget, delay, retries, user-agent, proxy, and raw-payload controls are fixed internally and are not public input fields.

{ "query": "recruitment", "maxItems": 3, "maxPages": 2 }

Output

Portable fields include jobId, title, companyName, locationText, descriptionText, postedAt, applyUrl, url, source, and scrapedAt. Rich fields preserve organization/location/salary structures, cleaned article HTML and text, overview data, tables, FAQs, dates, action links, breadcrumbs, and listing context. Promotional widgets, ad markup, scripts, and styles are removed from article content.

Requests are limited to official FreeJobAlert HTTPS hosts, use ordinary direct sequential access with fixed internal limits, and never use login, private APIs, or access-control bypasses. Records preserve public article tables, FAQs, eligibility, fees, dates, vacancy information, action links, structured metadata, receipts, field coverage, and direct-only quality evidence. The platform summary status is SUCCEEDED or FAILED; resultStatus separately reports COMPLETE, LIMITED, SKIPPED, DEFERRED, or FAILED. A confirmed access boundary on an official HTTPS FreeJobAlert response produces SUCCEEDED/SKIPPED; uncertain transport, HTTP 407, off-site redirects, and other unverified/non-2xx responses produce SUCCEEDED/DEFERRED. A zero-result parse is not enough to claim NO_DATA, so no NO_DATA outcome is emitted without explicit source evidence. Buffered records are discarded when access is blocked or cannot be verified. If usable Dataset rows were already persisted before a later failure, the platform status remains SUCCEEDED and the result is LIMITED. Diagnostics and candidate skips are stored only in key-value records, never in the Dataset. OUTPUT and the run key-value records preserve pagination, discovery, receipts, saved items, skips, and diagnostics.