NCS Jobs Search Scraper

Extract source-backed public job records from India's National Career Service (NCS), using the public /job-listing experience and its first-party API. The Actor uses bounded direct HTTPS requests, verifies each job against the matching detail response, and redacts public contact details from descriptions and retained source snapshots.

Source and output

• Public site: https://ncs.gov.in/job-listing
• First-party API: https://api.ncs.gov.in/api/v1/job-posts/search and https://api.ncs.gov.in/api/jobs/detail?id={id}
• Search cards expose titles, organizations, descriptions, location, skills, employment type, experience, salary, vacancies, and posting data. Public detail responses are retained in a sanitized source snapshot.
• Every emitted record matches an official NCS detail ID and contains at least 21 distinct populated source-fact groups. The quality count excludes IDs, URLs, aliases, and duplicated representations of the same fact.
• Recruiter names, email addresses, phone numbers, and other contact fields are omitted or redacted.

Input

See .actor/input_schema.json. Unknown fields are rejected. Detail enrichment is mandatory; retry, proxy, cookie, fingerprint, fixture, and debug controls are not public inputs.

Example:

{
  "mode": "search",
  "query": "nurse",
  "location": "",
  "maxItems": 3,
  "maxCandidates": 12,
  "maxPages": 1,
  "maxDetailRequests": 12,
  "maxRequests": 13,
  "deadlineSeconds": 120,
  "timeoutMs": 20000
}

Supported modes: search, searchMultiple, single, multiple, and startUrls. Detail URL inputs must use the official HTTPS ncs.gov.in domain.

Access handling and validation

Each request is direct, sequential, and single-attempt. HTTP 401, 403, 429, 451, or a visible source-authored denial or challenge stops all further requests, discards buffered job rows, preserves receipts, and records SKIPPED. HTTP 407, redirects, and transport failures stop the route and record DEFERRED; they are not treated as proof that NCS blocked access. No proxy, cookies, alternate identity, fingerprint changes, or challenge solving are used.

The Actor writes RUN_SUMMARY, OUTPUT_SUMMARY, RUN_DIAGNOSTICS, RUN_METADATA, and RUN_HEALTH to the default key-value store. status reports the platform process result (SUCCEEDED or FAILED); resultStatus reports dataset completeness (COMPLETE, LIMITED, SKIPPED, DEFERRED, NO_DATA, or FAILED). Only complete job records are written to the dataset, while diagnostics remain in the key-value store. Run npm test, npm run check, and apify validate-schema for offline validation. Before apify run, preserve existing actor storage, confirm that a new actor-relative APIFY_LOCAL_STORAGE_DIR does not exist, and run with that isolated relative path and --resurrect. Never use --purge or reuse an existing dataset directory. Validate with npm run validate -- <isolated-storage-dir>.