ATS Job Scraper — Greenhouse, Lever, Ashby & Workday

Give it a company domain. Get back every open job, in one clean schema, with parsed salary.

No board tokens to hunt down. No cookies. No proxies. No login.

Input:   ["stripe.com", "ramp.com", "palantir.com"]
Output:  1,022 jobs — normalized, deduplicated, salary parsed

---

What you actually get

Real rows from a single run across four different ATS platforms — one schema, no per-platform handling on your side:

Job title	Company	Location	Level	Salary	ATS
Security Engineer, Cloud	ramp	New York, NY (HQ) +3	mid	$211,400–290,600/yr	ashby
Backend Software Engineer - Application Development	palantir	Palo Alto, CA	mid	$135,000–200,000/yr	lever
Account Executive, AI Startups (Hunter)	stripe	San Francisco +1	mid	null	greenhouse
Software Engineer, Linux Graphics	nvidia	US, CA, Santa Clara	mid	null	workday

Salary reads null when the company doesn't publish one — it is never guessed or filled in.

Straight to a spreadsheet

Export as CSV and nested fields flatten with / — no post-processing:

title,companySlug,department,locations/0,workplaceType,seniority,salary/min,salary/max,salary/currency,postedAt,source
"Security Engineer, Cloud",ramp,Engineering,"New York, NY (HQ)",hybrid,mid,211400,290600,USD,2026-04-07,ashby
Backend Software Engineer - Application Development,palantir,Dev,"Palo Alto, CA",hybrid,mid,135000,200000,USD,2026-06-18,lever
"Account Executive, AI Startups (Hunter)",stripe,Startups - Account Executives (NA),San Francisco,,mid,,,,2026-08-19,greenhouse
"Software Engineer, Linux Graphics",nvidia,,"US, CA, Santa Clara",,mid,,,,2026-08-25,workday

Full records also carry jobId, employmentType, applyUrl, descriptionText, descriptionHtml, contentHash and scrapedAt. Turn off includeDescription for a smaller, faster payload.

---

Why this one

Most ATS scrapers make you do the hard part: find the company's board slug, figure out which ATS they use, then run a different scraper for each one and reconcile four different output shapes.

This Actor does all of that for you.

	Typical ATS scraper	This Actor
Input	Board slug you must find yourself	Company domain
ATS coverage	One per Actor	Greenhouse + Lever + Ashby + Workday
Output shape	Different per ATS	One schema for all
Salary	Raw text, if any	Parsed {min, max, currency, period}
Change tracking	—	Built in (new / updated / closed)
Proxies	Often required	None

---

Change tracking — turn it into a hiring monitor

Set onlyNewJobs: true, schedule the Actor daily, and each run returns only what changed:

• added — jobs posted since the last run
• updated — title, salary, location or description changed
• removed — job closed or pulled (a signal in itself)

The first run stores a baseline. Every run after that is a diff.

Use it for:

• Sales intelligence — a company opening 12 engineering roles is a buying signal
• Recruiting — know about a role the hour it goes live
• Market research — track hiring velocity and comp by company, team and level
• Job boards — keep a live index without re-ingesting everything

---

Input

Field	Type	Default	Description
companies	array	—	Domains (stripe.com) or board URLs (jobs.lever.co/palantir)
onlyNewJobs	boolean	false	Return only added/changed jobs vs. the previous run
trackingKey	string	default	Track multiple company lists independently
includeRemovedJobs	boolean	false	Also emit jobs that disappeared
includeDescription	boolean	true	Include full description text and HTML
maxJobsPerCompany	integer	0	Cap per company (0 = no limit)
workdayFetchDetails	boolean	true	Workday needs one extra request per job for descriptions

Board URLs are accepted directly, so you can skip resolution when you already know the board:

{
  "companies": [
    "stripe.com",
    "jobs.lever.co/palantir",
    "https://nvidia.wd5.myworkdayjobs.com/NVIDIAExternalCareerSite"
  ]
}

---

Output

Every record has the same shape regardless of which ATS it came from:

{
  "jobId": "ashby:ramp:34413f8d-26bf-4bbc-8ade-eb309a0e2245",
  "source": "ashby",
  "companySlug": "ramp",
  "companyDomain": "ramp.com",
  "title": "Security Engineer, Cloud",
  "department": "Engineering",
  "team": "Backend",
  "locations": ["New York, NY (HQ)", "Remote (Canada)", "Remote (US)", "Miami, FL"],
  "workplaceType": "hybrid",
  "employmentType": "FullTime",
  "seniority": "mid",
  "salary": { "min": 211400, "max": 290600, "currency": "USD", "period": "year" },
  "descriptionText": "ABOUT RAMP\n\nRamp is building the smart infrastructure for finance teams…",
  "descriptionHtml": "…",
  "applyUrl": "https://jobs.ashbyhq.com/ramp/34413f8d-…/application",
  "postedAt": "2026-04-07T17:12:35.753Z",
  "scrapedAt": "2026-08-24T18:03:28.104Z",
  "contentHash": "593a0ec77ef05d3c",
  "changeType": "added"
}

contentHash changes only when the job's meaningful content changes — use it for your own diffing.

seniority is derived from the title: intern, junior, mid, senior, principal, manager, director, executive.

---

Salary parsing

Ashby exposes structured compensation, so those values are exact. For Greenhouse, Lever and Workday the range is parsed out of the posting text, handling the formats companies actually use:

$173,000.00 - $259,600.00/yr     →  173000–259600 USD/year
€55k–€70k per year               →  55000–70000 EUR/year
USD 90000 to 120000 annually     →  90000–120000 USD/year
The salary range is $10,000/month →  10000 USD/month

Version numbers, headcounts and date ranges are not mistaken for salary.

Coverage measured against live boards:

Board	Jobs	Postings stating salary	Parsed
Ashby / ramp	137	137	137 (100%)
Lever / palantir	308	234	234 (100%)
Greenhouse / databricks	824	450	444 (99%)

Companies that don't publish salary return salary: null. Nothing is invented.

---

Speed

No browser, no proxies — just the public JSON endpoints each ATS already serves.

Board	Jobs	Time
Greenhouse / databricks	824	0.5 s
Greenhouse / stripe	577	0.7 s
Lever / palantir	308	1.4 s
Ashby / ramp	137	0.3 s

Workday is slower by design: its list endpoint returns no descriptions, so each job needs one extra request.

Set workdayFetchDetails: false to skip that — much faster and cheaper, but descriptions come back empty and jobs open in several offices read "2 Locations" instead of naming them. With the default (true) you get the full description and the real location.

---

Supported ATS

ATS	Input accepted
Greenhouse	domain, boards.greenhouse.io/<slug>, job-boards.greenhouse.io/<slug>
Lever	domain, jobs.lever.co/<slug>
Ashby	domain, jobs.ashbyhq.com/<slug>
Workday	domain, <tenant>.wd*.myworkdayjobs.com/<site>

More ATS platforms are being added. Need one in particular? Open an issue on the Actor and it gets prioritized.

---

Notes

• Only public job board data is collected — the same pages any visitor can open. No accounts, no logins, no personal candidate data.
• If a company's careers page can't be matched to a supported ATS, that company is skipped with a warning and the rest of the run continues.
• A company can have more than one board. The one with the most open jobs is used as primary; the others are still reported.