Career Site Job Scraper

Scrape job listings from any company's career site. Give it a domain like vercel.com — it automatically detects which applicant tracking system that company uses and returns every open role in one normalized schema. No board slugs, no per-ATS setup.

What does Career Site Job Scraper do?

It turns a list of company domains into a clean, structured job feed. Point it at 5 companies or 500 — it figures out which of 14 applicant tracking systems each one runs on, pulls every open posting, and normalizes the output so every source shares identical field names.

Click Start to try it with the default companies — no configuration needed. Then schedule it to run daily and get only what changed since the last run.

Sample output

This record is copied verbatim from a real run of this Actor (only descriptionHtml and descriptionText are shortened below — everything else, including the null fields, is exactly what the Actor returned):

{
  "sourceId": "ashby",
  "companySlug": "ramp",
  "companyName": null,
  "jobId": "34413f8d-26bf-4bbc-8ade-eb309a0e2245",
  "title": "Security Engineer, Cloud",
  "descriptionHtml": "<h1><strong>About Ramp</strong></h1><p>Ramp is building the smart infrastructure for finance teams, embedded in the transaction flow of every dollar a business spends...</p>",
  "descriptionText": "ABOUT RAMP\n\nRamp is building the smart infrastructure for finance teams, embedded in the transaction flow of every dollar a business spends. We automate how over $200B in annualized spend flows in and out of 70,000+ companies...",
  "locationRaw": "New York, NY (HQ)",
  "locations": [
    { "raw": "New York, NY (HQ)", "city": null, "region": null, "country": null },
    { "raw": "Remote (Canada)", "city": null, "region": null, "country": null },
    { "raw": "Remote (US)", "city": null, "region": null, "country": null },
    { "raw": "Miami, FL", "city": null, "region": null, "country": null }
  ],
  "workplaceType": "hybrid",
  "department": "Engineering",
  "team": "Backend",
  "employmentType": "FullTime",
  "seniority": null,
  "compensation": {
    "currency": "USD",
    "minAmount": 211400,
    "maxAmount": 290600,
    "interval": "year",
    "raw": "$211.4K – $290.6K • Offers Equity"
  },
  "applyUrl": "https://jobs.ashbyhq.com/ramp/34413f8d-26bf-4bbc-8ade-eb309a0e2245/application",
  "jobUrl": "https://jobs.ashbyhq.com/ramp/34413f8d-26bf-4bbc-8ade-eb309a0e2245",
  "publishedAt": "2026-04-07T17:12:35.753Z",
  "updatedAt": null,
  "requisitionId": null,
  "scrapedAt": "2026-08-09T19:02:29.589Z",
  "contentHash": "bc292d54781897269d7a38c2bf2efc4b621b51977c70e0d138cd6b20847a10ff"
}

Every record has the same shape regardless of which ATS it came from. Some fields are null by design rather than by error — for example Ashby's public API doesn't expose a company name or requisition ID, so those stay null for Ashby records but are populated for sources that provide them. compensation is structured (not just a text string) when the source publishes salary data — Ashby is the most common source for this.

How to use Career Site Job Scraper

1. Click Start to run it with the sample companies, or replace them with your own.
2. Paste the company websites you care about into Company domains — one per line, plain domains like vercel.com. You do not need to know their ATS. Already have board links? Paste those instead — https://jobs.lever.co/spotify, https://boards.greenhouse.io/anthropic and the other twelve platforms are recognised and fetched directly, with no lookup step.
3. Leave Return only new and changed jobs on if you plan to schedule it. Then set a daily schedule and each run returns only the delta. In a verified test run, a second pass over 205 existing postings with this option on returned zero records — every one was correctly skipped as unchanged.
4. Get results from the dataset, the API, or push them straight into Google Sheets, Slack, Zapier, Make or n8n.

Input configuration

The only field most people touch is Company domains. Everything else has a working default.

Field	What it does	Default
Company domains	Websites to track, or board links if you have them	ramp.com, vercel.com
Return only new and changed jobs	Skips postings identical to the previous run	On
Include job descriptions	Full description text and HTML	On
Max jobs per company	Cost cap. 0 means unlimited.	0
Direct ATS sources	Skip auto-detection when you already know the slug	Empty
Proxy configuration	Not needed for these sources	Off

Pricing

How much does it cost to scrape company job listings?

You pay per job returned, plus a small fee per company whose ATS we resolve. Nothing else.

Event	Price
Job scraped	$0.006
Company ATS detected	$0.002

Worked example: tracking 50 companies averaging 40 open roles each costs 50 × $0.002 + 2,000 × $0.006 = $12.10 for the first full run.

With Return only new and changed jobs switched on, a daily follow-up run typically returns only a small fraction of that volume, since unchanged postings are never charged for the job-scraped event. This is not a hypothetical: a verified run of 205 postings against this Actor, repeated with the same input, returned zero new or changed records on the second pass.

Output & integrations

Every record is returned in a stable schema with these fields: sourceId, companySlug, companyName, jobId, title, descriptionHtml, descriptionText, locationRaw, locations, workplaceType, department, team, employmentType, seniority, compensation, applyUrl, jobUrl, publishedAt, updatedAt, requisitionId, scrapedAt, contentHash.

Export as JSON, CSV, Excel or XML, or read the dataset through the Apify API. Works with the standard Apify integrations: Google Sheets, Slack, Zapier, Make, n8n, and webhooks.

contentHash is a fingerprint of the posting's content — use it to detect edits without diffing every field yourself.

Frequently asked questions

Is this Actor affiliated with any of the supported platforms? No. This is an unofficial tool and is not affiliated with, endorsed by, or connected to any of those companies. All product names are trademarks of their respective owners.

Is this legal? It reads the same public job board endpoints that companies publish so their openings can be distributed and indexed. It does not access private data, and it does not require or use any login. You are responsible for how you use the data, including compliance with applicable data protection law.

Why not just use each ATS's API separately? You would need fourteen integrations with fourteen sets of field names and fourteen slug formats - one returns XML, one needs a POST body and pagination, one only exposes its jobs inside a JSON block embedded in a web page - and still no way to go from a company domain to the right board. This does all of it in one call and gives you one schema.

Which ATS platforms are supported? Fourteen: Greenhouse, Lever, Ashby, Workable, SmartRecruiters, Recruitee, Personio, BambooHR, Teamtailor, Pinpoint, Rippling, Workday, Comeet and JOIN.

Five of them do not publish job descriptions on their public listing endpoint — BambooHR, Rippling, Workday, JOIN and SmartRecruiters return title, location, department and the apply link, but the description only exists on the posting page itself. Those records come back with descriptionText: null rather than with a fabricated summary.

Two platforms cannot be auto-resolved from a domain, because their address is not derivable from a company name:

• Workday needs a tenant, a data centre number and a free-text site name (nvidia.wd5.myworkdayjobs.com/.../NVIDIAExternalCareerSite).
• Comeet needs an account id and a separate token.

For both, a link on the company's careers page is picked up automatically. Otherwise pass them through Direct ATS sources as tenant:wdN:site and uid:token.

How many of my companies will be resolved? Two different things decide this, and it helps to keep them apart.

1. Does the company use a supported ATS? Fourteen platforms are supported (listed below). Companies on anything else - Gem, iCIMS, Comeet, or an in-house system - cannot be returned at all. On samples of tech companies picked before their ATS was known, this accounted for nearly every miss.

2. If they do, will we find it? This is the part we control, and it is high: 43 of the 45 companies that had a supported board were resolved - about 19 out of 20.

So the share of your list that comes back depends mostly on which platforms your companies use. On a list of AI and infrastructure companies it was around 7 in 10; on a list of developer tool companies, closer to 1 in 2.

Resolution works in two steps. First it reads the HTML your company's careers page serves and looks for a link to a supported board. Many modern marketing sites render their listings in the browser, so that HTML contains nothing to find. When that happens, the Actor derives likely board names from the domain and queries each ATS directly, which skips the page entirely.

A derived board name can belong to a different company that happens to share a name, so every derived match has to prove it belongs to you: the board must reference your domain or link back to your website. If neither holds, the company is reported as unresolved along with the board name we saw, and its jobs are not returned - handing you another company's jobs would be worse than handing you none. If you know the match is right, pass it through Direct ATS sources.

You are never charged for a company that is not resolved.

A company I added returned nothing. Why? Detection reads the HTML a company's careers page serves and looks for a link to one of the five supported boards. Two cases return nothing:

• The company uses an ATS not listed above.
• The careers page loads its listings in the browser with JavaScript, so the HTML the server sends contains no board link at all. This is common on marketing sites built as single-page apps, and no amount of retrying will surface a link that is not in the served HTML.

In both cases, use Direct ATS sources to supply the ATS and slug yourself — the run skips detection for that company and fetches the board directly. Detection is a convenience, not a requirement.

Companies that cannot be detected still cost a little run time (detection is capped at roughly 15 seconds per domain) but are never charged for, since company-detected only fires on a successful match.

Do I need a proxy? No. These sources are public JSON endpoints and work without one.