Workday Job Board Scraper

Scrapes open jobs from any company's Workday careers site — title, full location list, department, full description. No login, no API key.

Thousands of enterprises run their careers site on Workday (NVIDIA, Salesforce, and many more), all reachable through the same public JSON API the site's own frontend calls. Useful for talent-market research, competitor hiring tracking, and sales lead generation from which teams a company is actively scaling.

What makes this different

Multi-location jobs are actually resolved. Workday's list endpoint collapses a job posted to five offices into the string "5 Locations", with no way to expand it there — the real list only exists on the per-job detail endpoint. This Actor fetches it and gives you the genuine site list in locations, not a placeholder you'd have to re-query for yourself.

Page-size traps are handled invisibly. Workday hard-caps its page size at 20 and returns a bare, unhelpful HTTP 400 for anything larger — no error message pointing at the actual cause. This Actor paginates at the size Workday actually accepts and you never see the 400.

Tenant shard discovery. Workday splits companies across wd1–wd12+ hosts, and the shard isn't guessable from the company name. This Actor probes for the right one automatically, so you only need the tenant id and site id from the company's own careers URL — not which numbered shard they happen to sit on.

A wrong site id fails with the actual cause. Workday returns the identical generic error for "wrong site" and "wrong shard host," which would otherwise send a naive client cycling through all twelve hosts for no reason. This Actor recognizes Workday's specific "site not found" response and reports it immediately with the exact URL to check.

Input

Field	Type	Notes
tenant	string	From the careers URL, before .wdN.myworkdayjobs.com.
site	string	The path segment after the domain.
searchText	string	Optional filter, same as the site's own search box.
includeDetail	boolean	Full description + resolved locations, one extra request per job.
maxItems	integer	0 = every match.
wdHost	string	Advanced: skip shard auto-detection.
incremental	boolean	Only jobs not seen in previous runs.
proxyConfiguration	object	Not normally needed.

Finding tenant and site

From a URL like https://nvidia.wd5.myworkdayjobs.com/NVIDIAExternalCareerSite: tenant is nvidia, site is NVIDIAExternalCareerSite. Not every company uses Workday, and a wrong site id fails clearly for that run rather than returning something silently empty.

The cost of full detail

With includeDetail on, this Actor makes one request per job in addition to the list pages — that's what recovers the description and the resolved location list, which the list endpoint alone cannot provide. For a very large board, consider maxItems or searchText to scope the pull, or turn detail off for a fast title/location-only pass.

Output

{
  "job_id": "myworkdayjobs.com:Director--Rack-Scale-Software-Architecture_JR2012499",
  "title": "Director, Rack Scale Software Architecture",
  "location": "5 Locations",
  "locations": ["US, CA, Santa Clara", "US, TX, Austin", "..."],
  "time_type": "Full time",
  "hiring_organization": "NVIDIA Corporation",
  "posted_label": "Posted 30+ Days Ago",
  "url": "https://nvidia.wd5.myworkdayjobs.com/NVIDIAExternalCareerSite/job/..."
}

Each run also writes a RUN_COVERAGE record to the key-value store with what was requested, what came back, and any failure — so a partial run is visible rather than silent.

Local development

pip install -r requirements.txt
python test_local.py nvidia NVIDIAExternalCareerSite --max 25 --out sample_output.json
python test_local.py salesforce External_Career_Site --search python --no-detail

sample_output.json is real output from a live run against NVIDIA's board.