Expired Jobs API for Advanced LinkedIn Job Search API

What this Actor does

Observe explicit closed-to-applications status on actual native LinkedIn job-view pages. The Actor requires the native job ID and title plus the exact header label “No longer accepting applications”. A phrase found only in a job description, an active page, unrelated profile or missing document is not a closed-status result.

The keyless default uses Exa's public MCP content-search API. Every request uses search type auto and maxAgeHours=0. This is the provider's documented always-live-crawl contract. Cached fallback and cache-only search types are never used. No vendor API key, source login, generated cookie, browser identity change or scraping proxy is required.

Freshness is supplied by this provider contract. The public tool does not expose individual crawl timestamps or native HTTP statuses; the Actor explicitly records directSourceNetworkVerified=false and sourceCrawlTimestampKnown=false. It does not pretend that a direct native fetch or an exact crawl timestamp was independently verified.

Scope and defaults

This is a current closed-status snapshot, replacing the previous credential-only rolling expired-ID feed as the default. It does not establish when a job expired, whether it was closed by the poster, or whether an employer filled the vacancy. Older postings can be observed closed today. exactExpirationTimeKnown and sourceVacancyFilledStatusKnown remain false. timeRange must be all; recent-expiration windows are rejected rather than invented.

Example input:

{
  "searchQuery": "Data Analyst",
  "maxItems": 10,
  "timeRange": "all",
  "sampleMode": false
}

A nonempty startUrls list restricts output to specific native LinkedIn job IDs. The default searches for Data Analyst postings with an explicit native closed header. maxItems is a cap, not a promise to fabricate missing records. The Actor deduplicates by native job ID and fails when no qualifying actual source document is available.

Output and provenance

Results contain native jobId, jobTitle, jobUrl, the combined source employer/location header, observedClosedAt and qualified closed status. observedClosedAt is when the provider response was retrieved, not an invented expiry date. No exact expiration timestamp is emitted. Every record includes its actual provider request ID, source-text SHA-256, native header evidence and the forced-fresh policy. Missing optional fields are omitted, not replaced with nulls, placeholders or synthetic values.

Two anonymized examples illustrate the output shape only:

[
  {
    "jobId": "1234567890",
    "jobTitle": "Example Analyst",
    "jobUrl": "https://www.linkedin.com/jobs/view/1234567890",
    "expirationStatus": "closed_to_applications",
    "exactExpirationTimeKnown": false,
    "dataMode": "live_public_search_recrawl"
  },
  {
    "jobId": "2345678901",
    "jobTitle": "Example Engineer",
    "jobUrl": "https://www.linkedin.com/jobs/view/2345678901",
    "expirationStatus": "closed_to_applications",
    "exactExpirationTimeKnown": false,
    "dataMode": "live_public_search_recrawl"
  }
]

Limits and failure handling

The public provider can impose rate or availability limits. Failed requests, provider errors, missing text, source mismatches and no-result responses fail the run; no stale fallback is substituted. Requests and response sizes are bounded, and source HTML or text is never executed as code.

Inspect SOURCE_METADATA for the actual request, forced freshness policy and returned source documents. RUN_SUMMARY reports actual counts. ERROR_DIAGNOSTIC is separate from the dataset. Local CLI evidence and Apify production datasets are distinct. Maintenance should be cleared only after genuine nonempty output has passed current local and production verification.

Provider references: Exa Contents freshness controls and the official public MCP implementation. Apify execution and configured result-event charges apply in the cloud. Public-provider availability is not guaranteed.