Greenhouse, Lever & Ashby Jobs Scraper

Scrape job postings from any company running Greenhouse, Lever or Ashby — three of the most widely deployed applicant tracking systems — and get them back in one schema, whichever ATS they came from.

Built on the providers' own public JSON APIs. No proxies, no login, no browser, no anti-bot workarounds.

What makes this different

You don't need to know which ATS a company uses. That is the whole point. Give it stripe and it works out that Stripe is on Greenhouse; give it ramp and it finds Ashby. Single-ATS Actors push that lookup back onto you, which is unhelpful when you have 200 companies and no idea how they're split.

One schema across all three. Greenhouse calls a title title, Lever calls it text. Lever timestamps in epoch milliseconds, Ashby in ISO 8601, Greenhouse in ISO with an offset. Ashby says FullTime, Lever says Full-time. All of it is normalized before you see it.

Salary, including where the provider doesn't publish it. Ashby returns structured compensation. Greenhouse and Lever return none at all — any pay range lives in the posting prose, because US pay-transparency laws require it in the text rather than in a field. This Actor reads both, and tells you which is which:

"salary": {
  "min": 211400, "max": 290600, "currency": "USD",
  "interval": "year", "source": "structured", "raw": "$211.4K – $290.6K"
}

source is structured (from the provider, reliable) or parsed (read from prose, best-effort). raw is the exact text the numbers came from, so you can audit any parse you don't trust. When no range can be established confidently, salary is null — the parser is deliberately biased toward returning nothing rather than guessing, because a wrong salary silently corrupts your dataset in a way an empty field never does.

A bad slug doesn't cost you the run. Unrecognized companies are recorded in the COMPANIES key-value record with a reason, and the other boards continue.

Input

Give it board slugs, full board URLs, or a mix:

{
  "companies": ["stripe", "https://jobs.lever.co/palantir", "https://jobs.ashbyhq.com/ramp"],
  "titleKeywords": ["engineer"],
  "locationKeywords": ["remote", "berlin"],
  "remoteOnly": false,
  "includeSalary": true
}

Pasting a board URL is worth it — the ATS is read straight off the hostname, which skips detection entirely.

Field	Default	Notes
companies	—	Required. Slugs (stripe) or board URLs.
provider	auto	Pin to one ATS only if every company on the list uses it.
titleKeywords	[]	Case-insensitive substring; any match keeps the job.
locationKeywords	[]	Matches secondary locations too.
departmentKeywords	[]	Matches department or team.
remoteOnly	false	Inferred from location text on Greenhouse.
employmentTypes	[]	⚠️ Greenhouse publishes none — this filter excludes all Greenhouse postings.
postedAfter	null	ISO date. Excludes undated postings.
includeDescription	false	Adds 5–15 KB per job. Salary parsing works without it.
includeSalary	true
maxJobsPerCompany / maxJobsTotal	0	0 = no limit. Applied after filtering.
concurrency	3	Boards fetched at once.

Filters are applied before you are charged. If you ask for engineering roles in Berlin, you pay for those — not for the other 540 postings that had to be fetched to find them.

Output

One record per job:

{
  "provider": "ashby",
  "company": "ramp",
  "companyName": "ramp",
  "id": "34413f8d-26bf-4bbc-8ade-eb309a0e2245",
  "title": "Security Engineer, Cloud",
  "department": "Engineering",
  "team": "Backend",
  "location": "New York, NY (HQ)",
  "locations": ["New York, NY (HQ)", "Remote (Canada)", "Remote (US)"],
  "workplaceType": "hybrid",
  "isRemote": true,
  "employmentType": "full-time",
  "employmentTypeRaw": "FullTime",
  "salary": { "min": 211400, "max": 290600, "currency": "USD", "interval": "year", "source": "structured", "raw": "$211.4K – $290.6K" },
  "publishedAt": "2026-04-07T17:12:35.753Z",
  "updatedAt": "2026-04-07T17:12:35.753Z",
  "url": "https://jobs.ashbyhq.com/ramp/34413f8d",
  "applyUrl": "https://jobs.ashbyhq.com/ramp/34413f8d/application",
  "scrapedAt": "2026-08-17T09:00:00.000Z"
}

Two extra records are written to the key-value store: SUMMARY (counts by provider, how many carried salary, how many were remote) and COMPANIES (per-board outcome, including why any board failed).

What this Actor will not do

• No LinkedIn, Indeed or Glassdoor. Those are login-walled or Cloudflare-defended, and scraping them means a permanent proxy bill and an anti-bot arms race. This Actor reads official public APIs only, which is why it doesn't break.
• No SmartRecruiters. Its robots.txt is Disallow: / for every agent but LinkedIn's. Greenhouse, Lever and Ashby all permit the paths used here.
• No unlisted postings. Ashby's isListed: false marks jobs a company deliberately hid from its public board; those are dropped.

Field coverage by provider

Not every provider publishes every field. Rather than fabricate them, the Actor reports null:

Field	Greenhouse	Lever	Ashby
Title, location, URL, dates	✅	✅	✅
Department / team	✅	✅	✅
Employment type	❌ never published	✅	✅
Workplace type	⚠️ inferred from location	✅	✅
Structured salary	❌ parsed from text	⚠️ rare	✅
Company display name	✅	❌ slug used	❌ slug used