ATS Detector: Careers Page and Job Board Finder

Which ATS does a company use? For sales teams and recruiting tool builders: find careers page, job board URL, proof. $3.00 per 1,000 boards found, misses free. ATS lookup of 14 applicant tracking systems: detect Greenhouse, Lever, Workday, Ashby. Reads no postings; hiring-signals counts roles. Try it: run the prefilled input (https://www.figma.com) and open the ATS and job boards found table. An ATS detector and careers page finder.

Price (pay per event): $0.003 per job board found ($3.00 per 1,000 boards found), plus the platform's $0.00005 start event per run. Sites with no board, blocked sites and errors have no per-site charge.

What it does not do: Does not read or count job postings: give board_url to hiring-signals for that.

Quickstart (with your Apify token):

• API: curl -X POST "https://api.apify.com/v2/acts/edgef-4~careers-page-detector/run-sync-get-dataset-items" -H "Authorization: Bearer $APIFY_TOKEN" -H "Content-Type: application/json" -d '{"urls":["figma.com"]}'
• MCP call-actor: {"actor":"edgef-4/careers-page-detector","input":{"urls":["figma.com"]}}
• Python: ApifyClient(os.environ["APIFY_TOKEN"]).actor("edgef-4/careers-page-detector").call(run_input={"urls":["figma.com"]})

What it does

Give it company sites; for each it finds the careers page and the hosted job board behind it. It names the ATS platform (the applicant tracking system) and the board, and quotes the exact link, frame or script address that proves it. For Greenhouse, Lever, Ashby and SmartRecruiters it also returns a ready input for a job board reader.

Who uses it

• Sales and research teams who turn a list of company sites into a list of job boards to watch.
• Recruiting tool builders who need to know which applicant tracking platform a company uses.
• Analysts who measure platform share across a set of companies.

Input

Field	Type	Default	What it does
urls	list of text	required	Company sites or domains, for example example.com or https://www.example.com/

Output

One record per site. Sample record (SYNTHETIC: example.com data, not a real company):

{
  "id": "site:www.example.com",
  "site": "https://www.example.com/",
  "careers_url": "https://www.example.com/careers/",
  "platform": "greenhouse",
  "board_slug": "examplecorp",
  "board_url": "https://boards.greenhouse.io/examplecorp",
  "evidence": "https://boards.greenhouse.io/examplecorp/jobs/4000001",
  "evidence_kind": "link",
  "confidence": "high",
  "reader_input": {"platform": "greenhouse", "slug": "examplecorp"},
  "source_url": "https://www.example.com/careers/",
  "fetched_at": "2026-10-07T18:00:00Z",
  "status": "ok",
  "error": null,
  "error_detail": null
}

Field rules: evidence is the address exactly as written in the page at source_url. confidence is high for a link or a frame and medium for a script. careers_url is the link that clearly means the company's own jobs: a link text such as Careers, Jobs, Join us or We are hiring, or a careers word as a whole part of the path or of the host name. A product page about career sites is not taken. When links about careers were seen but none is clear, careers_url stays empty and the not_found record has confidence low. A site where no platform address was found has status not_found; blocked means the site did not let the Actor read the homepage, and error says which case it is: robots_disallow when its robots.txt disallows the homepage, blocked when the site refused the request with HTTP 401, 403 or 429 or with a bot challenge (error_detail then starts with "the site refused the request" and names the answer); an address that is not a public site, or a site that cannot be read, is error with the reason. A robots.txt that answers with a 4xx status other than 429 counts as no rules (RFC 9309). A robots.txt that cannot be read (a 5xx status, a 429, a timeout, a network failure or a redirect to another site) gives status error with error robots_unreadable, and the site is not read; it is never called a disallow. A host name that does not resolve gives status error with error dns_failure.

Platforms (14): Greenhouse, Lever, Ashby, SmartRecruiters, Workable, Recruitee, Personio, Teamtailor, BambooHR, JazzHR, Breezy HR, Workday, Jobvite, Pinpoint.

What a run costs

The price is $3.00 per 1,000 sites where a platform was found, plus the platform's start fee of $0.00005 per run (charged once, because the Actor runs with at most 1 GB of memory). Checking 1,000 sites in which 600 boards are found costs $1.80005. Sites without a detection, blocked sites and errors are not charged.

Limits and what it will not do

• Per site it reads the homepage and, when it is on the same host or its www form, the careers page, each with its redirects inside those two hosts (at most 5 hops), plus robots.txt once for each of the two hosts it actually requests: three requests for a site without redirects, five when the homepage redirects to the www form. robots.txt is asked before every request and every redirect hop. A careers page on another host is recorded in careers_url and never fetched.
• It matches link, frame and script addresses only. A page that names a platform in its text is not a detection.
• A service or marketing host under a platform's domain (such as www, landing, support, docs, academy or trust) is never reported as a board. On the site of a platform vendor itself, links to its platform domain count as its own pages: only its careers link can name its board.
• A site that redirects to another domain is reported as an error, because only the site's own host and its www form are requested.
• It follows robots.txt and never works around a block. No login, no cookies, no proxies, no browser: a board drawn by scripts after the page loads is found only through its script address.
• It identifies itself as edge-careers-page-detector/0.2 in every request. A site owner can block it in robots.txt with User-agent: edge-careers-page-detector and Disallow: /. It never pretends to be a browser.
• It does not read the job postings themselves, and it collects no data about people.

Source and terms

• The sites you list are read as any visitor would, with robots.txt obeyed and at most 2 requests per second per site. Check that the sites you list allow this.
• Platform addresses are matched by host and path patterns of the public board pages of each platform.
• This Actor is not affiliated with any of the 14 platforms or with the companies whose sites it reads.

Questions and answers

Why not_found on a site I know uses a platform? The careers page may load the board with a script the Actor does not run, or link it from a page deeper than the careers page. The careers_url field shows where it looked.

What is reader_input? An item you can paste into the boards list of a job board reader that reads Greenhouse, Lever, Ashby and SmartRecruiters.

Does it follow the careers link to another domain? No. It records the address and matches it against the platform rules without requesting it.

Which languages does it understand for the careers link? English, German, French, Spanish, Italian, Dutch, Portuguese, Polish, Finnish and Swedish link texts and paths.

Is a script detection as good as a link? It is marked medium: the script proves the platform and the board name, but not that the page lists open roles.