Greenhouse Jobs Scraper — Company Career Pages, No API Key

Pull every open role from any Greenhouse job board by company token. No API key, no login, no cookies, no proxy, no browser. You give it stripe, it gives you all 572 of Stripe's open postings with titles, departments, offices, full descriptions and a direct apply link for each one.

Greenhouse powers the careers page of a very large share of funded startups and mid-market tech companies. Every one of those boards is served by a public JSON endpoint that Greenhouse documents and keeps open. This Actor reads that endpoint properly, normalises the payload and hands you clean rows.

---

What you get, with the fill rate actually measured

We ran this Actor across 9 real boards — 2,628 live postings (stripe, anthropic, databricks, airbnb, coinbase, discord, robinhood, figma, dropbox) and counted how often each field came back populated. These are the real numbers, not a wish list:

Field	Fill rate	Notes
title	100%
companyName	100%	As Greenhouse has it, e.g. "Stripe"
jobId	100%	Greenhouse's public posting id
internalJobId	100%	Greenhouse's internal id
location	100%	The single location string on the posting
locations[]	100%	Every office attached to the posting
department / departments[]	100%	Requires content=true, which we always send
postedAt	100%	first_published, ISO 8601
updatedAt	100%	ISO 8601
jobUrl	100%	Canonical posting URL
applyUrl	100%	Deep link straight into the application form
boardUrl	100%
descriptionText	100%	Full description, HTML stripped and entities decoded
descriptionHtml	on request	Off by default; roughly triples dataset size
isRemote	100%	Always present as true/false — see the honesty note below
requisitionId	78.2%	Placeholder junk is nulled, see below
workplaceType	25.1%	Only when the location text actually says so
salaryMin / salaryMax / salaryCurrency / salaryText	57.0%	Parsed from the posting text, see below
educationRequirement	~98%	Greenhouse's education field
applicationDeadline	rare	Almost always null; most boards do not set it

Three fields this Actor deliberately does not return

The Greenhouse board API simply has no such data, on any board we measured. Rather than ship you three columns that are null on 100% of rows, they are absent from the schema:

• employmentType — Greenhouse's public board payload has no full-time/part-time/contract field.
• country — there is no ISO country on the posting. Use location and locations[].
• team — Greenhouse exposes a flat department list, not a department/team hierarchy.

If you need employment type or a normalised country, the Ashby Jobs Scraper and Lever Jobs Scraper in this same fleet both return them at ~99–100% fill, because those platforms actually publish them.

---

The two fields you should read carefully

Salary: 57%, and it is parsed, not published

Greenhouse's public board API exposes no structured pay field. We checked pay_input_ranges across all 2,628 postings on 9 boards: populated on zero of them. Anyone claiming a structured Greenhouse salary field is not reading the same API.

What we do instead is parse the posting text, because a lot of companies now write the range into the description to satisfy pay-transparency laws. That gets a range on 57.0% of postings. We then sanity-check every hit: of 1,207 extractions in a 2,245-row audit, 1,193 (98.8%) landed in a plausible annual band, the rest being genuine hourly or part-time roles.

Every salary row carries salarySource: "description" so you always know it was parsed rather than published. salaryText keeps the original snippet so you can verify it yourself. European formatting is handled: €92.300 — €130.000 is read as 92300–130000, not 92.3–130.

Remote: honest nulls instead of confident guesses

isRemote is always present. remoteSource tells you where the signal came from:

• inferred_location — 27% of rows. The location text says so: "US-Remote", "Poland - Remote OR Romania - Remote", "Remote in the US", "SF, NYC, remote". These are reliable.
• unknown — 73% of rows, returned as isRemote: false.

An earlier build also scanned the description body for the word "remote". We killed that path after measuring it: on 2,245 Stripe/Databricks/Airbnb postings it flagged 95 roles as remote whose location was Singapore, Dublin or Bengaluru. It was matching company boilerplate ("we are a remote-friendly company"), which describes the employer, not the job. A confident wrong true is worse than an honest false, so that inference is gone.

Filter on remoteSource === "inferred_location" if you only want rows where the board really said it.

Requisition IDs: junk removed

requisition_id comes back populated on 100% of postings, but 572 of 2,245 (25%) contained the literal placeholder text "See Opening ID". Those are nulled out. Real fill after cleaning: 78.2%.

---

Input

The only required field is the list of companies.

{
  "companies": ["stripe", "anthropic", "databricks"],
  "maxItems": 1000,
  "titleKeywords": ["engineer", "data"],
  "locationKeywords": ["New York", "Remote"],
  "remoteOnly": false,
  "includeDescription": true
}

Accepted company formats

All of these resolve to the same board:

stripe
https://boards.greenhouse.io/stripe
https://job-boards.greenhouse.io/stripe
https://job-boards.greenhouse.io/stripe/jobs/8130725
https://boards-api.greenhouse.io/v1/boards/stripe/jobs
https://anything.com/careers?for=stripe

You can also paste a newline- or comma-separated blob into a single string; it is split for you.

Input aliases

Different tools name things differently, so the Actor accepts synonyms and you never have to guess:

• companies: companies, company, boards, boardTokens, companyUrls, startUrls
• limit: maxItems, maxResults, limit, maxJobs
• keywords: titleKeywords, keywords, searchTitle
• locations: locationKeywords, locations, location
• departments: departmentKeywords, departments, department
• descriptions: includeDescription, includeContent, fullDescription

Filters

Option	What it does
maxItems	Hard ceiling on rows written, across all boards. This is your spend cap.
maxJobsPerCompany	Stops one 800-posting board from eating the whole budget
titleKeywords	Keep only titles containing one of these
locationKeywords	Match against location and every entry in locations[]
departmentKeywords	Match against the department list
remoteOnly	Keep only isRemote: true
postedAfter	ISO date; drops anything published earlier
includeDescription	Off makes the dataset far smaller
includeHtmlDescription	Adds raw HTML; off by default
dedupe	Drops repeat ids, and repeats of company+title+location
concurrency	Boards fetched in parallel, 1–15, default 5

---

Output sample

{
  "source": "greenhouse",
  "companyToken": "stripe",
  "companyName": "Stripe",
  "jobId": "8130725",
  "internalJobId": 3520748,
  "requisitionId": null,
  "title": "Account Executive, AI Startups (Hunter)",
  "department": "1653 Startups - Account Executives (NA)",
  "departments": ["1653 Startups - Account Executives (NA)"],
  "location": "San Francisco",
  "locations": ["US", "San Francisco"],
  "isRemote": false,
  "workplaceType": null,
  "remoteSource": "unknown",
  "salaryMin": null,
  "salaryMax": null,
  "salarySource": null,
  "postedAt": "2026-08-19T14:02:07-04:00",
  "updatedAt": "2026-08-19T14:02:07-04:00",
  "jobUrl": "https://stripe.com/jobs/search?gh_jid=8130725",
  "applyUrl": "https://job-boards.greenhouse.io/stripe/jobs/8130725#app",
  "boardUrl": "https://job-boards.greenhouse.io/stripe",
  "descriptionText": "Who we are About Stripe ...",
  "scrapedAt": "2026-08-22T04:11:02.884Z"
}

Four saved dataset views ship with the Actor: Job overview, Compensation, Remote & locations and Apply links, so you can export a clean CSV without picking columns by hand.

---

How errors are handled

Errors never enter your dataset. Charging you for a row that says "this failed" is charging you for a message. Everything that did not work goes into the FAILURES record in the run's key-value store:

{
  "runFailed": false,
  "boardsRequested": 3,
  "boardsWithJobs": 2,
  "jobsDelivered": 815,
  "duplicatesDropped": 0,
  "byReason": { "board_not_found": 1 },
  "entries": [
    { "target": "notarealboard", "reason": "board_not_found",
      "detail": "no Greenhouse board with this token (HTTP 404)",
      "at": "2026-08-22T04:11:02.884Z" }
  ]
}

The reasons are deliberately distinct, because they mean different things:

• board_not_found — HTTP 404. That company is not on Greenhouse, or the token is wrong. This is the normal outcome when you feed a list of companies, and it is not a bug.
• board_empty — HTTP 200 with an empty list. The board is real; the company just has nothing open right now.
• unresolvable_input — we could not derive a token from what you passed.
• budget_exhausted — maxItems was reached before this board was written. Nothing is ever skipped silently.
• fetch_failed — network or upstream error after retries.

If nothing was delivered, the run ends FAILED. A run that quietly succeeds with an empty dataset is how you find out a week later that your pipeline has been dead. You get a clear failure and a reason.

---

Pricing and the free tier

Pay per event, and only for results:

• Actor start — one small charge per gigabyte of memory
• Job scraped — charged after the row is written to the dataset

Rows are pushed in batches and the per-job event is only charged for rows that actually landed. Nothing is billed for a failure, a filtered-out posting or a deduplicated repeat.

The free tier returns real data. There is no proxy requirement, no API key, no credential of any kind. Everything this Actor touches is a public JSON endpoint. A free Apify account running {"companies": ["stripe"], "maxItems": 10} gets ten real jobs. Nothing in the code path throws because you lack a paid feature.

Use maxItems as your budget control: it is a hard ceiling enforced before rows are written.

---

Speed and cost efficiency

Greenhouse serves one board per request, so a board of 800 postings costs exactly one HTTP call. In our measured run, 2,628 postings from 9 boards took 5.0 seconds. Requests are gzip-compressed, retried with exponential backoff on 429/5xx, and never retried on 404.

Note that content=true is not optional and we always send it. Without it Greenhouse omits content, departments and offices from the response entirely — that is why some scrapers return no department data at all. The cost is a bigger payload (Stripe's board is 360 KB without descriptions, 4.4 MB with) which gzip absorbs.

If you are crawling 50+ large boards in one run, raise memory to 2 GB.

---

Common questions

Do I need a Greenhouse API key? No. The board endpoint is public and unauthenticated. The Harvest API needs a key; this is not that.

How do I find a company's board token? It is the path segment in their Greenhouse URL. If their careers page is job-boards.greenhouse.io/acme, the token is acme. Many companies embed the board on their own domain — view the page source and look for boards-api.greenhouse.io/v1/boards/<token> or a gh_jid parameter.

Why does jobUrl sometimes point to the company's own site? Greenhouse returns absolute_url, which large companies redirect to their branded careers page. applyUrl always points at the Greenhouse application form itself, so use that one for automation.

Can I get application questions or the form schema? Not in this Actor. It returns postings.

Is there an EU board host? Yes, and boards.eu.greenhouse.io/<token> URLs are parsed correctly.

How fresh is the data? Live. Every run hits Greenhouse directly; nothing is cached.

---

Related Actors in this fleet

• Ashby Jobs Scraper — structured salary ranges on ~72% of postings, plus a real remote flag
• Lever Jobs Scraper — team, commitment, workplace type and country at ~99–100%
• Workday Jobs Scraper — the enterprise side, by career-site URL
• Startup Jobs Aggregator — give it company handles, it finds each on Greenhouse, Ashby or Lever and returns one deduplicated schema

Legal

This Actor reads a public, unauthenticated JSON endpoint that Greenhouse publishes so job boards and aggregators can syndicate postings. It sends no credentials, solves no challenges and bypasses no access control. You are responsible for how you use the data, including any applicable data-protection rules.