Career Site Job Board Scraper — Greenhouse, Lever, Ashby & 4 more ATS

Paste a list of companies. Get back every open role on their job board, read live at the moment you run it, with a judged one-line verdict on each posting.

No database of yesterday's jobs. No 5,000-job ceiling. No API key.

in  →  ["https://boards.greenhouse.io/dropbox", "https://jobs.lever.co/binance"]
out →  one row per open role  +  one hiring-momentum summary per company

---

Why this one

Most ATS scrapers on the store hand you a cached copy of someone else's crawl, capped at a few thousand rows per run, and then point you at a paid API when you need more. This Actor calls the company's own job board at run time, so:

	This Actor
Freshness	Read live, per run. A role posted an hour ago is in your first run.
Per-run job limit	None. A 824-posting board comes back whole.
Coverage	Seven boards — Greenhouse, Lever, Ashby, Workable, SmartRecruiters, Personio and Recruitee — auto-detected from a slug or URL.
Judgement	Every row is scored, not just copied — see below.
Movement	Hand it the dataset of an earlier run and each row says whether it is new, each company says how its open-role count moved and which roles are gone.

The judgement layer

Other scrapers give you fields. This one gives you a verdict:

• headline — the row in one line: Senior Backend Engineer — Remote (United States), posted 3d ago, $180k–220k
• freshness — fresh (≤30d), aging (≤90d), stale (older)
• postedAgeDays — how long this role has actually been open
• ghostJobRisk — none / low / medium / high, with the reasons in ghostJobFlags: open over 90 or 180 days, never updated, or a duplicate title sitting on the same board
• seniority — intern / junior / senior / principal / manager / director / executive, read from the title
• salary — min, max, currency parsed out of whatever the board published
• remote and workplaceType — normalised across four boards that each say it differently
• keywordHits — which of your keywords this role matched

And once per company, a company-summary row: department and seniority mix, remote share, average posting age, top locations, and a hiring-momentum verdict — hiring hard, steady, slowing or stalled, based on how many roles opened in the last 30 days.

---

Input

Field	Type	Description
companies	array	Company slugs or careers-page URLs. All seven boards are detected automatically.
company	string	Alternative to the list — a single company.
keywords	array	Keep only roles whose title contains one of these words. Empty = every role.
locations	array	Keep only roles whose location mentions one of these (Remote works). Empty = everywhere.
postedWithinDays	integer	Keep only roles first published within N days. 0 = no date filter.
maxJobsPerCompany	integer	Ceiling per company. 0 = unlimited, which is the default.
includeDescription	boolean	Add the full job description as plain text. Off by default — descriptions are large.
monitorMode	boolean	Return only what changed since your previous run (default false). See Monitor mode below.
monitorKey	string	Optional name for the watch, e.g. competitor-hiring-weekly. Empty = derived from your input.
includeUnchanged	boolean	Monitor mode only — also return roles that are still open and unchanged (default false).
previousDatasetIds	array	Dataset IDs of earlier runs, used as the baseline for the change fields. Empty = comparisonStatus: no_baseline.
disclaimerAck	boolean	Acknowledges how this Actor sources data (see Data sources & responsibility).

Accepted address formats:

dropbox                                      (bare slug — probed across all seven boards)
https://boards.greenhouse.io/dropbox
https://job-boards.greenhouse.io/stripe/jobs/12345
https://jobs.lever.co/binance
https://jobs.ashbyhq.com/linear
https://apply.workable.com/blueground
https://jobs.smartrecruiters.com/Bosch
https://teeturtle.recruitee.com
https://personio.jobs.personio.de

A URL is always faster and unambiguous. A bare slug costs up to seven probe requests.

Example

{
  "companies": [
    "https://boards.greenhouse.io/databricks",
    "https://jobs.ashbyhq.com/linear"
  ],
  "keywords": ["engineer", "designer"],
  "locations": ["Remote", "United States"],
  "postedWithinDays": 30,
  "maxJobsPerCompany": 0,
  "includeDescription": false,
  "disclaimerAck": true
}

---

Output

One row per open role (rowType: "job"):

Field	Description
headline	The one-line verdict
company, companySlug, inputTarget	Who this role belongs to, and what you typed
ats, atsLabel	greenhouse / lever / ashby / workable
jobTitle	Role title
location, locations	Primary location and every listed location
remote, workplaceType	true/false/null, and remote / hybrid / onsite
department, departments	Team and department names
employmentType	Full-time, contract, intern, … where the board publishes it
postedAt, updatedAt	ISO dates from the board
postedAgeDays, updatedAgeDays	Age in days
freshness	fresh / aging / stale / unknown
ghostJobRisk, ghostJobFlags	Risk band and the reasons behind it
seniority	Inferred from the title
salary	{ min, max, currency, raw } when the posting publishes pay
keywordHits	Your keywords that matched
applyUrl	Direct application link
jobId, requisitionId	Board identifiers
jobDescription	Plain text, when includeDescription is on
sourceUrl, scrapedAt	Exactly which endpoint was read, and when
changeType, changeDetail	Monitor mode only — new / changed / removed / unchanged, and what moved

One row per company (rowType: "company-summary"): hiringMomentum, hiringMomentumReason, totalJobsOnBoard, jobsMatchingFilters, jobsDelivered, remoteJobs, remoteSharePercent, newLast30DaysPercent, averagePostingAgeDays, jobsWithPublishedSalary, departmentDistribution, seniorityDistribution, topLocations.

Rows with rowType: "notice" are status messages — an address with no board behind it, a company with no open roles, a limit reached. Notice rows are never charged.

Invalid input ends the run as SUCCEEDED with a notice row.

---

Monitor mode (weekly change alerts)

Hiring is a moving target. Turn on Monitor mode and the Actor compares this run with your previous run and returns only what moved on those boards:

changeType	Meaning
new	A role that was not on the board the last time you ran this — the one recruiters and job-seekers actually want
changed	The title or location moved; changeDetail shows what it said before
removed	No longer listed: filled, withdrawn or expired
unchanged	Still open, still identical — withheld by default, set includeUnchanged to receive it

Every run also returns one free change-report row, e.g. 12 new job postings, 3 gone, 1 changed since 2026-08-19, with newCount, removedCount, changedCount and unchangedCount. A quiet week returns just that row and no job rows.

The company summary row still describes the whole board (hiring momentum, department mix, remote share) so the change rows always have context.

Put it on a weekly schedule (3 steps)

1. Run the Actor once with Monitor mode on — the first run saves a baseline and returns every open role.
2. On the run's page click Actions ▸ Schedule Actor, and pick Every week.
3. Add an integration on the schedule (Slack, email, webhook) so new roles at the companies you watch reach you without opening Apify.

Runs are matched by monitor key. Leave monitorKey empty and it is derived from your input, so one schedule keeps one history automatically; set it (e.g. competitor-hiring-weekly) to keep several independent watches. State is kept in a named key-value store holding only job identifiers and a short title/location label — never descriptions.

A board that could not be read this run is left untouched: its roles are never reported as "removed" because of a network error.

---

Compare against an earlier run

Monitor mode keeps its own history. previousDatasetIds does the opposite: you choose the baseline, by pasting the dataset ID of any earlier run of this Actor. Nothing is stored between runs and nothing extra is charged — the comparison arrives as extra fields on rows you were already getting.

On every job row:

Field	Meaning
comparisonStatus	compared, compared_without_closures, baseline_unreadable, or no_baseline when you gave none
isNewSinceBaseline	The role was not in the baseline. null without a baseline.
previouslySeenAt	When the baseline last saw this role
postedAtChangedSinceBaseline	The board restamped the posting date — often a repost

On every company-summary row:

Field	Meaning
rolesOpenNow / rolesOpenInBaseline / rolesOpenChange	Open roles then, now, and the difference
newRolesSinceBaseline	How many of today's roles are new
closedSinceBaseline / closedRoles	How many are gone, and the first 25 of them with title, location and last-seen date
hiringChangeHeadline	The same thing in one line

Two rules keep this honest. Without a baseline there is no trend: comparisonStatus is no_baseline and every change field is null — today's role count is never dressed up as movement. And closures are only claimed when the run read whole boards: if keywords, locations, postedWithinDays or maxJobsPerCompany were set, a role missing from this run may simply have been filtered out, so the status is compared_without_closures and the closure fields stay null.

Pricing

Pay per event:

Event	Price	When
run-started	$0.005	Once per run, when your company list is accepted and the boards are read.
job-listed	$0.003	Once per job posting written to the dataset, with its full judged row.
company-summary	$0.01	Once per company, for the summary row. Only written when at least one job row was delivered for that company.

A 40-role company therefore costs 0.005 + 40 × 0.003 + 0.01 = $0.135.

Notice rows — unreadable board, no open roles, no matching filters, limit reached, and the monitor change report — are not charged.

In Monitor mode the same job-listed price applies to each delivered change row, removed rows included: a role that has just been filled is exactly the answer you scheduled the run for. A week with no movement delivers no job rows and therefore costs only the run-start fee.

---

Data sources & responsibility

This Actor reads the public, unauthenticated job-board endpoints that each applicant tracking system serves to a company's own careers page:

Greenhouse       https://boards-api.greenhouse.io/v1/boards/<slug>/jobs?content=true
Lever            https://api.lever.co/v0/postings/<slug>?mode=json
Ashby            https://api.ashbyhq.com/posting-api/job-board/<slug>?includeCompensation=true
Workable         https://apply.workable.com/api/v1/widget/accounts/<slug>?details=true
SmartRecruiters  https://api.smartrecruiters.com/v1/companies/<slug>/postings?limit=100
Recruitee        https://<slug>.recruitee.com/api/offers/
Personio         https://<slug>.jobs.personio.de/xml

• One request per company on five of the seven boards. SmartRecruiters pages its feed at 100 postings, so a larger board costs one request per page. SmartRecruiters and Personio omit the job description from the list feed, so turning on includeDescription adds one read of the public posting per role on those two boards only — leave it off and they cost one request like the rest.
• Two companies at a time, with an honest bot User-Agent (ntriq-ats-jobs/1.0).
• No login, no captcha, no anti-bot measure is bypassed. Nothing behind an authentication wall is read.
• Only job-posting data is returned. No recruiter names, no applicant data, no personal contact details.

This Actor is unofficial and is not affiliated with, endorsed by, or sponsored by Greenhouse, Lever, Ashby, Workable, SmartRecruiters, Personio, Recruitee, or any employer whose board you read. Those names appear only to say which public boards it reads. You are responsible for complying with the terms of service of each applicant tracking system and each employer, and for using the results lawfully. You are the data controller for everything you collect with it, including any obligations under GDPR or comparable law.

ghostJobRisk is a heuristic built from posting age, update dates and duplicate titles. It is a signal to investigate, not a statement about any employer's hiring intentions.

---

Notes

• Boards that publish no posting date are still returned; their freshness is unknown and postedWithinDays never filters them out.
• Ashby and Lever publish pay ranges more often than Greenhouse and Workable, so salary is populated unevenly. It is parsed from what the board actually published — nothing is estimated.
• A company whose board exists but has no open roles returns a free no-jobs notice, not an error.

Works well with

• whois-domain-lookup — Qualify the hiring company by its domain: registrar, age and registrant details.
• lighthouse-website-audit — Score the company site for performance and SEO before you pitch or apply.
• theharvester-osint — Collect public emails and hosts for the same company domain.

Every run of this Actor also returns these suggestions in the nextSteps field of its summary row.