SEEK Jobs Scraper — Australia & New Zealand Listings with Full Ad Bodies

Scrape job listings from SEEK AU and NZ: title, company, location, salary label, work type, classification, listing date — and optionally the complete ad body, expiry date, status and advertiser contact. HTTP-only, no account, no browser.

Input

Field	Type	Default	Description
searchQueries	array	required	One search per entry, e.g. registered nurse
country	string	AU	AU (seek.com.au) or NZ (seek.co.nz)
location	string	—	Optional place filter, e.g. Sydney. Also the way past the result ceiling
maxItemsPerQuery	integer	100	Jobs per search
includeDescription	boolean	false	Fetch the full ad body; one extra request per job

Output

{
  "_input": "python developer",
  "_source": "S1-jobsearch-api+S2-redux-island",
  "_scrapedAt": "2026-09-09T10:58:11Z",

  "jobId": "94305187",
  "url": "https://www.seek.com.au/job/94305187",
  "title": "Python Developer",
  "teaser": "If you are a Python Developer looking for a change…",
  "bulletPoints": ["…", "…"],

  "companyName": "Just Digital People",
  "advertiserId": "29028300", "advertiserName": "Just Digital People",
  "employerId": "6368", "employerName": "…", "employerUrl": "https://au.seek.com/companies/…",

  "location": "Brisbane QLD",
  "locations": ["Brisbane QLD"],
  "countryCode": "AU",

  "salaryLabel": "$87,790 - $112,607 per annum (pro rata)",
  "workTypes": ["Full time"],
  "workArrangements": ["On-site"],
  "classifications": [
    { "classification": "Information & Communication Technology",
      "subClassification": "Developers/Programmers" }
  ],

  "listingDate": "2026-08-31T04:33:55Z",
  "listingDateDisplay": "9d ago",
  "isFeatured": false,
  "logoUrl": "…",

  // only with includeDescription:
  "abstract": "Join the … team, delivering vital virtual care around Australia",
  "description": "At Amplar Health, we're transforming how healthcare is delivered…",
  "descriptionHtml": "<p>…</p>",
  "status": "Active", "isExpired": false, "isVerified": true, "isLinkOut": true,
  "expiresAt": "2026-10-09T03:44:11.000Z",
  "listedAtUtc": "2026-09-09T03:45:24.586Z",
  "shareLink": "https://au.seek.com/job/94514328?tracking=…",
  "contactMatches": [{ "type": "Phone", "value": "(03) 8622 5666" }]
}

Three things worth knowing

totalCount is not how many jobs you can get. A bare search reports 163,544 matches, but SEEK stops serving results after roughly 500 per search — pageSize=20 dies after page 27, pageSize=100 after page 5. The ceiling is on the result set, not the page size, so asking for bigger pages does not help. The actor logs a warning when a query matches more than it can reach; the fix is to narrow the search — add a location, or split one broad query into several specific ones.

The API names fields in the plural, and the singular forms are always null. locations, workTypes and classifications carry the data; location, workType and classification exist on the object and are permanently empty. This actor reads the plural ones and also emits a convenience location string holding the first label.

The ad body comes from GraphQL, not the job page. The page carries it in a window.SEEK_REDUX_DATA island, and that was the first implementation — but the HTML page sits behind an anti-bot layer that the API host does not: from Apify it answered 403 on every TLS profile through datacenter proxies, and roughly one request in six through residential, while /api/… and /graphql on the same host answered 200 throughout. The actor uses POST /graphql, which needs no proxy upgrade and moves ~7 KB instead of a 240 KB page. (The page also has no JobPosting JSON-LD — its only structured-data node is WebSite.)

Errors

_error	Meaning
invalid_input	Empty search query
no_results	The search ran and SEEK returned nothing
blocked	Every TLS profile was refused by the server
network_error	The ladder never reached the server — DNS, timeout, reset
unexpected_shape	200 with a structure this actor was not written against

A job whose body could not be fetched still produces its full search-level row, with _detailError naming the cause — detail_not_found (GraphQL has no job for that id, so the listing was removed or expired), network_error, detail_blocked, or detail_unexpected_shape. Dropping a real job because one extra request failed would be the worse outcome. The ad-body fetch repeats the whole TLS ladder once before giving up, since a page that fails on a truncated body usually succeeds seconds later.

If every query fails, the run itself fails rather than reporting success over an empty dataset.