Workday Jobs Scraper: Any Workday Careers Site, By Name

Every open role from any Workday careers site. Type a company name for one of the 2,898 employers this Actor has already verified, or paste any Workday URL. Titles, locations and links as clean JSON, no API key and no proxy. Workday is where large enterprises post: NVIDIA about 2,000 open roles, Salesforce about 1,500, Adobe about 750.

Features

• Type a company name, for 2,898 employers. The Actor ships a map of 3,911 career sites it has already read real postings from, so nvidia or pwc works without you hunting down a wd5.myworkdayjobs.com address. Anything outside that map still takes the URL, and is refused rather than guessed at.
• Any workday board, from the URL or token you already have. No API key, no login, no cookies, no proxy: this reads the public feed workday publishes so that job boards can index it.
• Every posting, not the first page. Paging is handled, including the places where workday reports a total it does not honour.
• One row per job, as clean JSON, with the same field names on every run.
• Filters that cost you nothing. Rows dropped by titleKeywords, locationKeywords or remoteOnly are never charged for.
• Errors are per company. One bad id does not end the run, and error rows are not charged.
• Run it on a schedule and the rows become a record of who started hiring and when.

How to use it

1. Click Try for free, or add this Actor to a task.
2. Put one or more board ids in the input. The example below is a real one.
3. Optionally narrow it with titleKeywords, locationKeywords or remoteOnly. Filtered rows are not billed.
4. Run it. Results appear in the dataset and can be exported as JSON, CSV, Excel or fetched from the API.

Input

Input	Type	Default	What it does
companies	array of strings	none	Company names, one per line, e.g. nvidia or pwc. Resolved against a shipped map of 2,898 Workday tenants this Actor has already read real postings from, and all of a company's career sites are read. A name that is not in the map is refused rather than guessed at, because Workday returns the same empty response for a company that does not exist and for one with no open roles, so a derived token could never be confirmed. Use boards for those
boards	array of strings	none	Workday careers page addresses, one per line, e.g. https://nvidia.wd5.myworkdayjobs.com/NVIDIAExternalCareerSite
titleKeywords	array of strings	none	Keep only roles whose title contains any of these, one per line, e.g. security or staff engineer
locationKeywords	array of strings	none	Keep only roles whose location contains any of these, one per line, e.g. London or Israel
remoteOnly	true or false	false	Keep only roles flagged remote by the ATS, or whose title or location says remote or anywhere
includeDescription	true or false	false	Fetch each job's own page for the full description, a REAL posting date, every office on a multi-location posting, and the employment type. One extra request per job, so it is off by default
maxResultsPerCompany	number	1000	Ceiling on roles taken from any single board
maxItems	number	none	A hard ceiling on rows for the entire run, across every company

{
  "boards": ["https://nvidia.wd5.myworkdayjobs.com/NVIDIAExternalCareerSite"],
  "titleKeywords": ["engineer"],
  "remoteOnly": true
}

Output

One row per job. This is the shape, with the fields workday actually publishes:

{
  "title": "Senior Software Engineer",
  "location": "Berlin, Germany",
  "url": "https://boards.example.com/jobs/8130725",
  "jobId": "8130725",
  "postedAt": "2026-08-20T09:14:02Z",
  "postedAtRelative": "Posted Today",
  "employmentType": "full-time",
  "isRemote": true,
  "company": "nvidia",
  "atsPlatform": "workday",
  "boardToken": "https://nvidia.wd5.myworkdayjobs.com/NVIDIAExternalCareerSite",
  "boardTokenResolved": "nvidia",
  "boardUrl": "https://nvidia.wd5.myworkdayjobs.com/NVIDIAExternalCareerSite",
  "seniority": "senior"
}

Field
title	the role as the company wrote it
location	as the board published it, EXCEPT where the board published a count instead of a place. Workday writes "2 Locations" into this field on multi-office postings, and 27% of some boards' rows carried a tally where a place belonged; those now carry the primary office read from the job's own URL, and locationCount keeps the number
url	the public posting, ready to open
jobId	the ATS's own id, stable across runs
postedAt	ISO 8601 UTC
postedAtRelative	a phrase, not a date: "Posted Today", "Posted 30+ Days Ago". This source publishes no timestamp, and a date invented from that phrase would read as exact when it is not
employmentType	full time, contract, intern, as the source says
isRemote	true only for genuinely remote roles
company	what you asked for, echoed back
atsPlatform	which system it came from
boardToken	the board id used
boardTokenResolved	the board id derived from your URL
boardUrl	the public board this came from
city	where the location names one
country	derived from the location, and from the job URL where the location is a phrase like "2 Locations"
region	state or province, where the location names one
seniority	derived from the title: intern, junior, senior, staff, lead, principal, director or executive. Empty where the title does not say, which is most of them
workArrangement	remote, hybrid or onsite
locationCount	how many offices one posting is open in, present only where the board published a count instead of a place
applicationDeadlineText	"3 days left to apply", on the few postings where the board publishes a closing date INSTEAD of a posting date. On those rows postedAtRelative is absent, because a deadline is not a posting time

What people use this for

Hiring data is not really about jobs. It is the earliest public signal a company gives that something changed, and it is why three different kinds of buyer end up on the same dataset:

• Sales and go-to-market. A company that opens six engineering roles this month is a company with new budget. Job postings say which team is growing and in which city, weeks before anything shows up in a funding announcement.
• Investors and market research. Headcount by function, tracked over time, across a whole portfolio or a whole sector. Every row carries the company, the team and the date, so a weekly run is a time series.
• Recruiting and talent. Where a competitor is hiring, which roles they have been trying to fill for months, and how fast a team is growing. Workday does not publish pay, so no row here claims to know it.

Run it once for a snapshot. Run it on a schedule and the same rows become a record of who started growing and when.

A company name, or the exact URL

For 2,898 employers you can just type the name:

{ "companies": ["nvidia", "pwc", "Wells Fargo", "Procter & Gamble", "Verizon", "Citi"] }

Names work the way people write them, not only the way Workday codes them: "Wells Fargo" finds the wf tenant, "Procter & Gamble" finds pg, "Morgan Stanley" finds ms, "The Walt Disney Company" finds disney.

Those are the tenants this Actor has already read real postings from, shipped with it as a map of 3,911 verified career sites. 524 of them publish on more than one site (PwC has six, GHR seven), and all of a company's sites are read, because somebody asking for PwC's jobs wants PwC's jobs and not whichever site happened to sort first. Names are matched loosely, so Hy-Vee, hy vee and hyvee.com all land on the same employer.

For anyone else, pass the careers URL:

{ "boards": ["https://nvidia.wd5.myworkdayjobs.com/NVIDIAExternalCareerSite"] }

A name that is not in that map is refused rather than guessed at. This is the one place this Actor will not try to be clever, and the reason is in the next section: a 200 from Workday does not prove a company exists, so a guess could never be confirmed, and a confident wrong answer is worse than being told to paste a URL. The map is not a guess, which is exactly why it is allowed: every entry in it is a site that has already returned real jobs.

What makes workday different

• These are the biggest boards on any system. Re-measured 2026-08-31: NVIDIA 2,000 open roles, Salesforce 1,515, Adobe 738. The counts move week to week, which is the point of scraping them rather than quoting them. Most large enterprises post here rather than on a startup ATS, so a jobs pipeline that stops at Greenhouse and Lever misses the companies people most want to watch.

• Workday caps the page size at 20 and answers 400 above it, undocumented, and that 400 reads as a malformed request rather than a paging problem. Measured: 20 returns 200, 50 is rejected.

• Workday's total lies after the first page. The real count arrives on page one and every page after it reports 0, so the obvious offset >= total loop is true immediately and stops at 40 of 2,000 postings, with no error and nothing in the log. This Actor keeps the first answer and stops only on an empty page.

• The board listing has no description, no real posting date and no office list, and includeDescription now goes and gets all three from the job's own page. This matters most for the date: the listing says "Posted 30+ Days Ago" and stops there, which is unbounded, so those rows carry no postedAt at all. A FedEx role reading "30+ days" was measured on 2026-08-31 to have been posted on 2026-05-05, which is 118 days. It is off by default because it costs one extra request per job.

• It needs the careers URL for any employer not in the shipped map of 2,898 verified tenants, and that is not laziness. A Workday address carries three separate unknowns: the host label, the data-centre number (wd1 up to wd504, and companies move between them) and the career-site name. Guessing across them would mostly bill you for 404s, so this asks for the page you already have.

Filters that cost you nothing

Filtered rows are not charged. Keywords match as plain text, so c++ and node.js mean exactly that rather than being read as regular expressions.

Integrations and API

Every run writes to a dataset you can export as JSON, CSV or Excel, or read from the Apify API. The Actor can be scheduled, called from another Actor, or wired into Make, Zapier, Slack, Google Sheets and the rest of Apify's integrations. It is also callable by an AI agent through the Apify MCP server, and the output schema means the agent gets field descriptions rather than raw JSON.

Looking for an Indeed or Glassdoor API? There isn't one, and this is why you don't need it

Indeed has no public jobs API. Neither does Glassdoor, ZipRecruiter or LinkedIn Jobs, and all four block you at the edge. Measured from an ordinary residential address on 2026-08-01, with normal browser headers:

indeed.com/jobs          403      0 bytes
glassdoor.com/Job/...    403      0 bytes
ziprecruiter.com         403      0 bytes
upwork.com/nx/search     403      0 bytes

Zero bytes. Cloudflare rejects the request before it reaches an application, so there is nothing to parse and no proxy budget that fixes it.

But none of those four originate job data. They aggregate it from company career pages, and those pages run on systems like workday that publish a free, keyless, public API, because companies want their openings indexed. That API answered with real jobs from the same connection, in the same minute.

Going to the source is also fresher. An aggregator shows you its last crawl. This shows you the board.

FAQ

Do I need an API key or an account with workday? No. This reads the public feed workday publishes for indexing. Nothing here is behind a login, a paywall or a bot wall.

What am I charged for? Rows returned. Rows removed by your filters are not charged, and neither are error rows.

Can I get only what changed since my last run? Run it on a schedule and compare jobId, which is stable across runs. For change tracking with the work already done, see the sibling Actors below.

A board came back empty. Is the company not hiring? On workday an empty board cannot be told apart from a wrong id, so the message says exactly that rather than claiming the company has nothing open.

Notes

• The source is a public API that companies publish deliberately. Nothing here is behind a login, a paywall or a bot wall.
• An empty board cannot be told apart from a wrong id on this system, and the message says so.
• Errors are per company. One bad token does not end the run, and error rows are not charged.

Related

• Oracle Recruiting Jobs Scraper: the same rows from Oracle Cloud careers sites, 1,066 employers by name, including Marriott, Kroger, Hilton and AutoZone
• ATS Jobs Search: search one index of jobs across many careers systems

Use it from an AI agent

This Actor is a ready-made tool for Claude, ChatGPT, Cursor or any other MCP client, through Apify's MCP server. Add it as a connector at:

https://mcp.apify.com?tools=glitchbound/workday-jobs-scraper

In Cursor or VS Code the same address goes in mcp.json:

{ "mcpServers": { "apify": { "url": "https://mcp.apify.com?tools=glitchbound/workday-jobs-scraper" } } }

The first call asks you to sign in to Apify, or takes an API token. From then on the agent sees this Actor's input fields and calls it directly.

Feedback

If this saved you time, a review on the Actor's Store page helps other people find it. If something is missing or wrong, the Issues tab on the same page is the fastest way to reach me, and it gets read.