Career Site Jobs Monitor — Greenhouse, Lever, Ashby, Workable, SmartRecruiters

Pull live job postings straight from company career sites. Give it a list of companies (an ATS board URL, an ats:slug, or just the company's careers page) and it returns every open role as flat JSON — with a stable dedupKey, ISO timestamps and filters, so you can run it on a schedule and get only the roles that are new since last time.

No login, no API key, no proxy. Each supported ATS publishes a public JSON endpoint that companies use to syndicate their own postings. This actor reads those endpoints and nothing else, which is why it's fast and doesn't break the way page scrapers do.

Supported ATS platforms

ATS	What you can pass	Descriptions
Greenhouse	greenhouse:stripe, https://boards.greenhouse.io/stripe, job-boards.greenhouse.io/…, EU boards	✔ (includeDescription)
Lever	lever:spotify, https://jobs.lever.co/spotify, EU boards	✔
Ashby	ashby:notion, https://jobs.ashbyhq.com/notion	✔ (+ published compensation)
Workable	workable:acme, https://apply.workable.com/acme, acme.workable.com	✖ (list endpoint only)
SmartRecruiters	smartrecruiters:BoschGroup, https://jobs.smartrecruiters.com/BoschGroup	✖ (list endpoint only)

Any careers page URL (https://acme.com/careers) also works: the actor fetches the page, finds the embedded board of any of the five ATSs, and continues from there. If a page uses an unsupported ATS you get a warning naming the URL, never a silent empty result.

What it does

1. Resolves each company entry to an ATS board (explicit, by URL, or by page discovery).
2. Reads the board's public JSON API, following pagination where the ATS paginates.
3. Normalises every posting into the same flat record regardless of ATS.
4. Applies your filters — title keywords, excluded keywords, locations (incl. remote), postedAfter.
5. Deduplicates within the run and pushes one record per posting, up to maxItems.

Input

Field	Type	Default	Notes
companies	array of strings	—	Required. ATS URLs, ats:slug, or careers page URLs
keywords	array	[]	Keep postings whose title contains any keyword (case-insensitive)
excludeKeywords	array	[]	Drop postings whose title contains any of these
locations	array	[]	Substring match on location; include "remote" to also keep remote-flagged roles
postedAfter	string	—	ISO date/date-time. Drops older postings; keeps postings with no date
includeDescription	boolean	false	Adds plain-text descriptionText (Greenhouse, Lever, Ashby)
maxItems	integer	500	Billing guardrail — hard cap on postings pushed per run
proxyConfiguration	object	off	Not needed for these APIs; enable only at very high volume

Example — every engineering role at four companies, UK or remote, posted this month:

{
  "companies": ["greenhouse:stripe", "https://jobs.lever.co/spotify", "https://jobs.ashbyhq.com/notion", "https://jobs.smartrecruiters.com/BoschGroup"],
  "keywords": ["engineer"],
  "locations": ["United Kingdom", "London", "remote"],
  "postedAfter": "2026-08-01",
  "maxItems": 200
}

Output

One JSON object per posting in the default dataset:

Field	Type	Description
source	string	greenhouse | lever | ashby | workable | smartrecruiters
company	string	Board slug on that ATS
companyName	string | null	Display name when the ATS provides it
jobId	string	Posting ID on the ATS
title	string	Job title
location	string | null	Location text
isRemote	boolean	ATS remote flag, or inferred from the location text
department, team	string | null	When the ATS provides them
employmentType	string | null	e.g. Full-time
compensation	string | null	Published pay range (Ashby)
postedAt	string | null	ISO 8601 publish date
updatedAt	string | null	ISO 8601 last update (Greenhouse)
url, applyUrl	string	Public posting / apply links
descriptionText	string | null	Plain text, only with includeDescription
dedupKey	string	Stable key source:company:jobId
scrapedAt	string	ISO 8601

Monitoring: new roles since last run

dedupKey never changes for a given posting, so the recipe is:

1. Schedule the actor (daily works well; hourly for hot accounts).
2. Either set postedAfter to the previous run's time, or diff each run's dataset against the last one on dedupKey — new keys are new roles, missing keys are closed roles.
3. Send the difference wherever you like (Slack, email, a Google Sheet) with an Apify integration.

Typical uses: hiring-signal alerts for sales and recruitment teams, tracking competitor hiring, feeding a niche job board, watching a shortlist of dream companies.

Pricing

Pay per event: one charge per posting pushed to the dataset. Postings removed by your filters are free. maxItems caps the spend of any single run.

Notes & limitations

• Workable and SmartRecruiters list endpoints don't include descriptions; includeDescription is ignored for them.
• A board that returns 404 (wrong slug, or the company moved to a different ATS) is reported in the log and skipped — it doesn't fail the run.
• Discovery reads the careers page HTML you give it. Boards injected by client-side JavaScript that never appear in the HTML source can't be found that way — pass the board URL directly.
• Sources are the ATSs' own public APIs; the actor sends a descriptive User-Agent and backs off on 429.

Development

npm install
npm test          # unit tests (parsers, normalisers, filters — no network)
apify run         # local run with storage/key_value_stores/default/INPUT.json