Career Site Jobs Scraper (Greenhouse, Lever, Ashby, Workday)

Read the open jobs of any company that hires through Greenhouse, Lever, Ashby, Workable, Recruitee, Personio, Workday, straight from the public job-board feeds these systems provide, and get them as clean, normalized JSON, CSV or Excel. One schema for every system: title, department, location(s), remote flag, employment type, posting date, salary when published, full description (text and HTML) and apply link.

Two modes. Companies mode: you list companies (career URLs, ats:id, or a company website / careers page) and get all their jobs. Discovery mode: leave Companies empty and the Actor scans the bundled company directory (biggest hiring boards first, maxCompanies default 200) and filters the jobs live by keyword, location, remote and date.

The directory: 9,670 company job boards, each checked live when the directory was built (4,140 Greenhouse, 278 Lever, 1,957 Ashby, 857 Workable, 850 Recruitee, 687 Personio, 901 Workday), with about 528,000 open jobs between them at that time. Companies not in the directory can always be read in Companies mode.

Who is it for

• Recruiters and sourcers who want a clean list of open roles at target companies, with location, department and salary.
• Job boards, newsletters and aggregators that need a fresh, structured feed taken from the employer's own career site.
• Market and talent analysts tracking who is hiring, for what, where and at what pay.
• AI agents and MCP users: every field is described in the dataset schema, so an agent can answer "which remote data jobs were posted this week" from the output.

What you get

One dataset item per job:

• jobId, ats, companyId, companyName
• title, department, team
• location, locations[], country, countryCode, isRemote, workplaceType (remote / hybrid / onsite), employmentType
• postedAt, updatedAt (ISO 8601, UTC)
• jobUrl, applyUrl
• descriptionText (clean text) and descriptionHtml
• salary with min, max, currency, period and the original raw text, when the employer published a range
• scrapedAt, sourceUrl, and an extras block with system-specific fields (requisition ids, departments, custom metadata, ...)

Only what the employer publishes in the job posting is returned. Names and e-mail addresses are not extracted into separate fields. Full field reference: ./SCHEMA.md and the Actor's dataset schema.

Input

Field	Description
companies	Career URLs, ats:id entries or company websites (see table below). Empty = Discovery mode
keywords	Keep jobs whose title contains any of these words (case-insensitive)
excludeKeywords	Drop jobs whose title contains any of these words
locations	Keep jobs whose location or country contains any of these texts; Remote also matches remote-flagged jobs
remoteOnly	Only jobs the employer marks as remote
postedWithinDays	Only jobs first published in the last N days (0 = no limit; jobs with no date are excluded when set)
ats	Discovery mode: restrict to these systems (greenhouse, lever, ashby, workable, recruitee, personio, workday)
includeDescription	Include description text and HTML (default true)
maxItems	Maximum jobs, newest first (default 20)
maxCompanies	Discovery mode: how many directory companies to scan (default 200)

Leaving the input empty reads one well-known company so the default run always works and returns 20 jobs in a few seconds.

Company entries

System	Careers URL	ats:id form
Greenhouse	https://boards.greenhouse.io/airbnb	greenhouse:airbnb
Lever	https://jobs.lever.co/palantir	lever:palantir
Ashby	https://jobs.ashbyhq.com/ramp	ashby:ramp
Workable	https://apply.workable.com/huggingface/	workable:huggingface
Recruitee	https://bunq.recruitee.com	recruitee:bunq
Personio	https://ottonova.jobs.personio.de	personio:ottonova
Workday	https://nvidia.wd5.myworkdayjobs.com/NVIDIAExternalCareerSite	workday:nvidia.wd5/NVIDIAExternalCareerSite

Also accepted: lever:eu:<slug> (EU Lever accounts), personio:com:<sub> (jobs.personio.com), workday:<tenant>.wd<N>/<site>, and a company website or careers page: the Actor fetches only that one page (robots.txt respected) and looks for an embedded Greenhouse, Lever, Ashby, Workable, Recruitee, Personio or Workday job board.

Input example (Companies mode)

{
  "companies": [
    "https://boards.greenhouse.io/airbnb",
    "lever:palantir",
    "https://jobs.ashbyhq.com/ramp"
  ],
  "keywords": [
    "engineer"
  ],
  "locations": [
    "Germany",
    "Remote"
  ],
  "maxItems": 100
}

Input example (Discovery mode)

{
  "companies": [],
  "keywords": [
    "data scientist"
  ],
  "remoteOnly": true,
  "postedWithinDays": 14,
  "ats": [
    "greenhouse",
    "lever",
    "ashby"
  ],
  "maxCompanies": 100,
  "maxItems": 200
}

Output example

{
  "schemaVersion": "1.0",
  "jobId": "8306554",
  "ats": "greenhouse",
  "companyId": "airbnb",
  "companyName": "Airbnb",
  "title": "AirCover Specialist, Claims Experience",
  "department": "Community Support",
  "team": null,
  "location": "Canada",
  "locations": [
    "Canada"
  ],
  "country": null,
  "countryCode": null,
  "isRemote": true,
  "workplaceType": "remote",
  "employmentType": null,
  "postedAt": "2026-10-28T14:52:05+00:00",
  "updatedAt": "2026-10-28T14:52:05+00:00",
  "jobUrl": "https://careers.airbnb.com/positions/8306554?gh_jid=8306554",
  "applyUrl": "https://careers.airbnb.com/positions/8306554?gh_jid=8306554",
  "descriptionText": "Airbnb was born in 2007 when two hosts welcomed three guests to their San Francisco home, and has since grown to over 5 million hosts who have welcomed over 2 b...",
  "descriptionHtml": "<div class=\"content-intro\"><p><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">Airbnb was born in 2007 when two hosts welcomed three gu...",
  "salary": null,
  "scrapedAt": "2026-11-06T12:00:00+00:00",
  "sourceUrl": "https://boards-api.greenhouse.io/v1/boards/airbnb/jobs/8306554",
  "extras": {
    "requisitionId": "ONE",
    "internalJobId": 3591434,
    "language": "en",
    "offices": [
      "Canada"
    ],
    "departments": [
      "Community Support"
    ],
    "metadata": {
      "Is this job part of ACC?": false,
      "Workplace Type": "Remote"
    }
  }
}

Use cases

• Talent sourcing: every open engineering role at 50 target companies, refreshed daily.
• Job aggregation: a deduplicated feed of jobs taken from the employer's own site (stable job URLs and apply links).
• Hiring-trend research: count open roles by department, country or remote policy over time (schedule the Actor and compare datasets).
• Pay transparency: collect published salary ranges by role and country.
• AI agents (MCP): ask in plain language, get structured jobs back.

Data sources

• Greenhouse: Greenhouse Job Board API (boards-api.greenhouse.io), the public API Greenhouse offers so companies can show their jobs on their own website. Its documentation states that GET endpoints need no authentication.
• Lever: Lever Postings API (api.lever.co/v0/postings), the public API Lever documents for building job sites. Published postings are also visible on each company's jobs.lever.co page.
• Ashby: Ashby public job posting API (api.ashbyhq.com/posting-api/job-board), the documented public endpoint behind Ashby-hosted job boards, called with compensation included.
• Workable: Workable public widget API (apply.workable.com/api/v1/widget/accounts/<subdomain>), the endpoint behind Workable's embeddable jobs widget.
• Recruitee: Recruitee public careers API (<company>.recruitee.com/api/offers/), the JSON feed behind Recruitee career sites.
• Personio: Personio public XML feed (<company>.jobs.personio.de/xml, or .jobs.personio.com), the feed Personio provides for career pages.
• Workday: the JSON endpoint of Workday career sites (<tenant>.wd<N>.myworkdayjobs.com/wday/cxs/...), the same one the public career site calls. It is not a documented API: the Actor checks robots.txt of each host before every request and uses it only on hosts that allow it.

All of these are public endpoints that need no account, key or login. The Actor identifies itself honestly with the User-Agent NexusScrapers-CareerSiteJobs/0.1 (+https://nexus-tools-dev.github.io/), reads robots.txt of every host (including Crawl-delay), limits concurrency per host, and backs off and retries on HTTP 429/5xx, then skips the company with a warning. It uses no proxies, no browser and no captcha handling. The employers' postings remain their property; you are responsible for using the data in line with the employer's terms and applicable law (for example, do not use postings to contact individuals).

Limits and good to know

• Greenhouse: salary is filled only when the company publishes a pay range on the board (common for US postings because of pay-transparency laws, rarer elsewhere). Up to 4 parallel requests per host.
• Lever: Salary ranges are returned only for postings where the company published one. robots.txt of Lever asks for Crawl-delay: 1: the Actor sends at most one request per second to Lever, one at a time, so big lists of Lever companies take proportionally longer.
• Ashby: Compensation is returned only when the company chose to show it on the posting. Up to 4 parallel requests per host. Large boards (a few MB of JSON) are processed one at a time and only matching jobs are kept.
• Workable: The widget API does not expose salary or an update date. Up to 4 parallel requests per host.
• Recruitee: Salary ranges are returned only when the company filled them in. Up to 4 parallel requests per host.
• Personio: The feed has no salary, no company display name (derived from the subdomain) and no remote flag (remote is detected only when the title or office says so). Up to 4 parallel requests per host.
• Workday: Some Workday tenants answer HTTP 422 to this request and are skipped with a warning. With includeDescription off, postedAt is approximate (derived from the 'Posted N days ago' label). One request at a time per Workday host, 0.3 s apart. The list gives 20 jobs per request and no description, so descriptions cost one more request per job: they are fetched only for jobs that pass your title/location filters, at most maxItems per company.
• Company websites: only the single page you give is read. If that page builds its job list with JavaScript and does not mention the job board in its HTML (many do not), no board is found: give the job-board URL instead.
• Jobs are returned newest first within the companies scanned; in Discovery mode each company contributes at most max(10, maxItems/4) jobs, so results come from several companies, and scanning stops once enough matching jobs are collected.
• locations and keywords are plain text matches on what the employer wrote (no translation, no synonyms).
• A company that does not exist, has no open jobs or blocks access is skipped with a warning in the log; the run still succeeds if at least one company was read.

FAQ

Do I need an API key or a proxy? No. These are the public feeds the systems provide for career pages.

Is this legal? The Actor only reads publicly available job postings through public endpoints that allow anonymous access, respects robots.txt and rate limits, and does not collect personal data. This is not legal advice; check your own use case.

How fresh is the data? Live at run time: the feeds are read when the Actor runs. postedAt is the date the employer published the job.

Why is salary often empty? Only some employers publish ranges. When they do, you get min, max, currency and period.

Can I get only new jobs since my last run? Not built in yet; schedule the Actor and compare jobId values between runs, or use postedWithinDays.

What happens at the budget limit? The Actor stops at maxItems or when your run's maximum charge is reached, and returns what it already collected.

Pricing

Pay per result: each job written to the dataset is one result. You are never charged for more results than maxItems or your run's spending limit allow. See the Actor's pricing tab for the current price.

API, MCP and AI agents

Call the Actor from the Apify API, the JavaScript or Python clients, or schedule it. The dataset schema documents every field, so MCP-enabled AI agents can use the output directly.