Company Jobs Scraper: Workday, Greenhouse, Lever & Ashby

You pay only for confirmed company matches.

Give it a list of companies and get back every open job they publish, in one clean table. Company Jobs Scraper reads each company's own job board at run time from nine applicant tracking systems (ATS): Greenhouse, Lever, Ashby, Workable, Recruitee, SmartRecruiters, Workday, Rippling and Gem. Use it as a jobs API for company career pages: run it daily in "new" mode and you get only the postings that appeared since the last run.

• Type a name, a domain or a URL. "Canva", zendesk.com and https://jobs.lever.co/spotify all work. Every row says how its company was matched (confidence, matchedVia). Only a board URL, or a board the website you gave links to, counts as confirmed and is charged. A match found from the name alone is returned free and marked medium.
• One schema for every ATS. Title, department, locations, remote/hybrid/onsite, employment type, salary range, posting date, job URL and apply URL use the same field names whatever the source.
• Salary where the employer publishes it. Pay ranges come from the ATS's structured fields, with currency and period (year, month, hour).
• New-since-last-run tracking. Each row carries changeType (new, updated, unchanged, removed). Pay only for the rows you ask for.
• Live data. Jobs are fetched from each board's API during your run; there is no database in between.

Who uses it

You are	You use it to
Sales or partnerships team	Watch target accounts for hiring signals: a company posting its first "Head of Security" role is a buying signal.
Recruiter or talent-intelligence analyst	Track competitors' openings, team growth and published salary bands.
Job board or newsletter builder	Pull fresh postings from a curated list of companies every day, with a stable id for de-duplication.
AI agent builder	Give an agent a tool that answers "Which of these companies is hiring ML engineers in Berlin right now?"

Supported job boards

ATS	Enter it as	Found from a company name?	Salary	Posting date
Greenhouse	Stripe, https://boards.greenhouse.io/stripe, greenhouse:stripe	Yes (published name checked)	When published	Yes
Lever	Spotify, https://jobs.lever.co/spotify, lever:spotify, EU: https://jobs.eu.lever.co/acme	Yes	When published	Yes
Ashby	Notion, https://jobs.ashbyhq.com/notion, ashby:notion	Yes	When published	Yes
Workable	Hugging Face, https://apply.workable.com/huggingface, workable:huggingface	Yes (published name checked)	No	Yes
Recruitee	bunq, https://bunq.recruitee.com, recruitee:bunq	Yes (published name checked)	When published	Yes
SmartRecruiters	ServiceNow, https://jobs.smartrecruiters.com/Canva, smartrecruiters:Canva	Yes (published name checked)	No	Yes
Workday	Salesforce, https://salesforce.wd12.myworkdayjobs.com/External_Career_Site, workday:salesforce.wd12/External_Career_Site	Yes	No	Day only, up to 30 days back
Rippling	Rippling, https://ats.rippling.com/rippling/jobs, rippling:rippling	Yes	No	No
Gem	Retool, https://jobs.gem.com/retool, gem:retool	Yes	No	Yes

In our test run on 2026-10-04 (25 real boards, 4,130 jobs), the share of jobs with a structured salary range was 71% on Ashby, 30% on Recruitee, 25% on Greenhouse, 15% on Lever and 0% on Workable. It depends entirely on what each employer chooses to publish.

SmartRecruiters, Workday and Rippling rows have title, location, workplace type (when given), job URL and posting date (except Rippling), but no description: their job lists do not include it.

In our test on 2026-10-08 with 90 real company names in three lists (one of them chosen by an outside tester), the actor matched 64 to the right board. All were found from the name alone, so all were returned free as medium. One more free medium match was a subsidiary's Workday site (ASML's Berlin site, whose main site is closed to crawlers). The other 25 were not found and not charged: they use systems this actor does not read (iCIMS, Oracle, Avature, Deel, in-house career sites), do not show their board on their careers page under a token their name gives, or block automated requests. Typing each company's domain instead of its name gave 22 confirmed and charged matches, where the company's careers page links the board. No charged match was wrong.

How to use it

1. Add companies to Companies, one per line: names, website domains, careers-board URLs or ats:token.
2. Optional: add Keywords (matched in titles), Locations, Remote jobs only, Only jobs with a published salary or Posted within (days).
3. Pick an Output mode: all for every open job, new for jobs not seen before in this watchlist, changes for new, updated and removed jobs.
4. Run it. Results are in the Jobs table; the Run summary shows which board each company resolved to, and why any entry was not found.
5. To get daily alerts, save the input as a task, set output mode to new and add a schedule. Connect the dataset to Google Sheets, Slack, Zapier, Make or a webhook through Apify integrations.

Input example

{
    "companies": ["Stripe", "huggingface.co", "https://jobs.lever.co/spotify", "ashby:ramp"],
    "keywords": ["engineer", "data"],
    "locations": ["Remote", "London"],
    "outputMode": "new",
    "watchlistName": "eng-roles",
    "descriptionFormat": "text",
    "maxJobs": 500
}

Output example

One row per job. This is a real posting from a public Lever board, with the description shortened:

{
    "id": "lever:zoox:f4746da4-8eb8-43e2-b7ce-bf3c7cf9640d",
    "ats": "lever",
    "boardToken": "zoox",
    "jobId": "f4746da4-8eb8-43e2-b7ce-bf3c7cf9640d",
    "company": "zoox",
    "title": "Autonomy System Test Engineer",
    "department": "Software",
    "team": "Software Platforms and Product",
    "location": "Foster City, CA",
    "locations": ["Foster City, CA"],
    "countryCode": "US",
    "workplaceType": "hybrid",
    "employmentType": "full_time",
    "employmentTypeRaw": "Full-time",
    "salary": { "min": 144000, "max": 193000, "currency": "USD", "period": "year", "text": null },
    "url": "https://jobs.lever.co/zoox/f4746da4-8eb8-43e2-b7ce-bf3c7cf9640d",
    "applyUrl": "https://jobs.lever.co/zoox/f4746da4-8eb8-43e2-b7ce-bf3c7cf9640d/apply",
    "publishedAt": "2026-05-04T23:11:01.125Z",
    "updatedAt": null,
    "descriptionText": "Autonomous vehicles have some of the largest, most complex software ever shipped in a safety-critical environment. Solving that problem is one of the most excit…",
    "changeType": "new",
    "firstSeenAt": "2026-10-04T11:20:05.120Z",
    "boardUrl": "https://jobs.lever.co/zoox",
    "confidence": "exact",
    "matchedVia": "url",
    "query": "https://jobs.lever.co/zoox",
    "scrapedAt": "2026-10-04T11:20:05.120Z"
}

company is the name the ATS publishes. Lever and Ashby publish none, so the actor uses the name you typed, or the board token when you gave a URL.

Run summary

The SUMMARY record in the run's key-value store lists every input with its status (resolved, not_found, invalid, duplicate, unsupported, error, skipped), the board used, confidence and matchedVia, other boards found under the same name, and counts of jobs on the board, matched, returned, new, updated and removed. totals.unconfirmed counts the companies returned free as medium.

How a company is matched

How a name or domain is looked up:

1. A domain or careers-page URL you give (zendesk.com, https://www.figma.com/careers): the actor reads that site's careers pages (the page you gave, /careers, and careers. and jobs. on the domain) and takes the job boards they link to. Pages the site asks crawlers not to read, and pages that block the request, are skipped.
2. Only when that finds no usable board, or for a plain name: the actor reads the careers pages of <name>.com and tries the name as a board token on each ATS ("Hugging Face" → huggingface, hugging-face). Both are guesses from the name, so they are weighed alike.

Every row and every summary entry carries:

• matchedVia: url (you gave the board URL or ats:token), site (the website you gave links to the board), name (found from the company name alone).
• confidence:
  • exact: you gave a board URL or ats:token. Charged.
  • high: the website you gave links to the board, and the ATS publishes no company name (Lever, Ashby, Workday, Rippling, Gem) or the same name as your domain's, ignoring case, spaces and suffixes such as "Inc." (Greenhouse, Workable, Recruitee, SmartRecruiters). Charged.
  • medium: nothing confirms the match. Either the board was found from the name alone, even under the same published name (unrelated companies share names: the Recruitee board "GONG" is a Polish ad agency, not Gong; a Workday tenant found by name may be a subsidiary's site), or the website you gave links to a board published under a name other than your domain's (toasttab.com links the board named "Toast"). Returned free. Give the board URL, or the company's domain when its careers page links the board, to confirm it.

A board is not used at all when it has no jobs or only jobs posted over a year ago (ATSs keep empty and abandoned accounts under many company names), or when it was found from the name and the ATS publishes another company's name, even a longer one that starts with yours ("Wise" does not match "Wise Worksite Field Sales"). When no board qualifies, the entry is not_found, is not charged, and lists the boards it skipped under alternatives. To watch a company with no openings right now, give its board URL.

Track only new jobs

Set Output mode to new and keep the same Watchlist name between runs. The actor remembers the jobs it has seen per watchlist and board in a named key-value store in your Apify account (jobboard-watchlists), so:

• The first run returns all current jobs, because nothing has been seen yet. For a baseline run, set Max jobs per company to 1: jobs beyond that cap are stored as seen (cost: about 1 row + 1 company check per company). Set it back to 0 for the following runs.
• Later runs return only jobs that were not on the board before.
• changes mode also returns jobs whose title, location, department, workplace type, employment type or salary changed (updated), and jobs that disappeared (removed). Description-only edits are ignored on purpose. Removed jobs are filtered by the title, location, workplace type, employment type and salary they had when last seen.
• If a run stops at Max jobs per run or at your spending limit, the new, updated and removed jobs it did not return come back on the next run.

How much does it cost?

This actor uses pay-per-event pricing. You pay for:

Event	Price	When it is charged
Job	$1.00 per 1,000 jobs ($0.001 each)	Each row from a confirmed match (exact or high) written to the dataset
Company checked	$0.50 per 1,000 companies ($0.0005 each)	Each confirmed company match. Unconfirmed (medium) and not-found companies are free.
Actor start	Apify's standard start fee	Once per run

Examples:

• 20 companies, all open jobs (about 2,000 rows): 20 × $0.0005 + 2,000 × $0.001 = about $2.01.
• 100 companies checked every day in new mode, about 40 new jobs a day: $0.05 + $0.04 = about $0.09 a day, $2.70 a month.

Use Max jobs per run and Apify's maximum cost per run to cap spending. The actor stops cleanly at either limit.

Use with AI agents (MCP)

This actor works as a tool for AI agents through the Apify MCP server. Each input field has a plain-language description, and the output uses stable field names.

• Add the actor to your MCP client (for example Claude or Cursor) through the Apify MCP server, then ask, for example: "Use Company Jobs Scraper to list remote data engineering jobs at Stripe, Ramp and Hugging Face posted in the last 7 days."
• Good agent inputs: companies (names or URLs), keywords, locations, remoteOnly, postedWithinDays, and descriptionFormat: "none" when the agent only needs titles and links (smaller, cheaper context).
• The run's status message gives a one-line result (for example "60 jobs from 4/5 companies (1 not found, 0 errors)"), and the SUMMARY record explains every miss so the agent can retry with a URL.

Call it from your code

curl -X POST "https://api.apify.com/v2/acts/jay_hong~jobboard-actor/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"companies": ["Stripe", "https://jobs.lever.co/spotify"], "keywords": ["engineer"], "descriptionFormat": "none"}'

Limits

• Only companies whose job board runs on one of the nine ATSs above. iCIMS, Oracle Recruiting, SuccessFactors, Taleo, Eightfold, Deel, LinkedIn and Indeed are not supported. There is no search across all companies on an ATS; you supply the companies.
• Workday: the first career site the tenant lets crawlers read is used (for example Salesforce's External_Career_Site, not its separate intern site); pass a site URL to pick another. If a company's main site is closed to crawlers, that may be a subsidiary's site, which is why a Workday tenant found by name alone is medium and free. Workday's search returns at most 2,000 jobs per career site. Workday career sites that ask crawlers to stay away are skipped (not_found, not charged).
• Some companies link their careers page to a board that has no public feed, or load the link only in the browser. Those need the board URL.
• Recruitee will require an employer token for its public API from 2027-02-10. From that date Recruitee entries return unsupported.
• Personio and Pinpoint boards are not supported at this time. Entries for them return unsupported and are not charged.
• EU-hosted Greenhouse boards (job-boards.eu.greenhouse.io) are not supported at this time. Entries for them return invalid and are not charged.
• Greenhouse, Workday and Rippling publish no employment type, so the Employment types filter drops their jobs. Workable does not distinguish hybrid from onsite.
• Name lookups try likely board tokens ("Hugging Face" → huggingface, hugging-face). Companies that use an unusual token and do not link their board from their careers page need the board URL.
• Data is what each employer publishes. Salary and locations are copied from the ATS fields. Remote status comes from the ATS's remote field, or from the word "remote" or "hybrid" in the location when the ATS has no such field (Greenhouse). Nothing is generated by AI.

FAQ

Where does the data come from? From each ATS's public job-board feed, the same one the employer's own careers page uses. Greenhouse, Lever, Ashby, Recruitee, SmartRecruiters and Gem document these APIs for building careers pages and job listings; Workable documents its public account endpoint in its help center. For Workday and Rippling there is no documented API: the actor reads the public JSON their career sites load in the browser. To find a board, the actor also reads the HTML of the company's careers pages. It does not log in, does not run a browser, and does not use proxies.

How fresh is the data? As fresh as the employer's board at the moment of your run. Nothing is served from a cache or a database.

Why did a company return "not_found"? It uses an ATS this actor does not support, its careers page does not link a supported board and it uses an unusual board token, or the only boards under its name had no recent jobs or publish a different company name (for example a longer legal name that only starts with yours). The summary lists which ATSs and tokens were tried and the boards that were skipped. Paste the board URL to fix it.

Can I get every job on Greenhouse or Lever? No. ATSs publish no directory of boards. You bring the list of companies.

Does it collect personal data? It returns job postings only. Fields such as the per-job application mailbox some ATSs expose are dropped. Job descriptions are employer text and can occasionally name a contact person; handle the data under the rules that apply to you (for example GDPR).

How fast is it? In our test run, 25 boards with 4,130 jobs took 29.6 seconds of fetching. A name lookup reads a few careers pages and tries several ATSs, so it takes longer than a board URL: 30 names took 100 to 170 seconds on 2026-10-08. Workday returns 20 jobs per request, so a site with 2,000 jobs takes about two minutes. The actor waits between requests to the same host (1 second for Lever and Workable), so very long lists take longer.

Is descriptionHtml safe to show on my site? It is the employer's HTML exactly as the ATS publishes it. Sanitize it before rendering it in a web page, as you would any third-party HTML.

Can I filter by salary amount? Turn on Only jobs with a published salary, then filter salary.min / salary.max in the dataset or your own code. Currencies and periods differ between jobs, so the actor does not convert them.

Changelog

• 0.1 (2026-10): first version. Five ATSs, name/domain/URL lookup, change detection, pay-per-event.
• 2026-10-08: a board found from the company name alone is no longer charged, even when the ATS publishes the same company name. Before, "Gong" was charged for the Recruitee board of GONG, a Polish ad agency. Only a board URL or ats:token, or a board the website you gave links to, is charged.
• 2026-10-08: name lookups use a board only when it has open jobs and, where the ATS publishes one, the same company name. Before, an empty Workable account or a board of a company whose name only started the same way could be returned as a match and charged.
• 2026-10-08: added SmartRecruiters, Workday, Rippling and Gem. Names and domains are looked up on the company's careers pages first. Boards whose jobs are all over a year old no longer count as a match.

Company Jobs Scraper is not affiliated with or endorsed by Greenhouse, Lever, Ashby, Workable, Recruitee, SmartRecruiters, Workday, Rippling or Gem.