Career Site Jobs Scraper: Greenhouse, Lever, Ashby & more

Give it a list of companies and get every open job from their own career sites, in one table with the same columns for every company. It works with Greenhouse, Lever, Ashby, Workable, SmartRecruiters, Recruitee and Personio.

You don't have to know which system a company uses. Enter stripe.com and the actor finds the job board itself.

Good for job boards and aggregators, recruiters tracking target companies, sales teams using hiring as a buying signal, and anyone building a job feed for an AI agent.

Input

• Companies: a domain (ramp.com), a careers page URL, or a job board URL (boards.greenhouse.io/discord, jobs.lever.co/palantir, jobs.ashbyhq.com/ramp).
• Optional filters: title keywords, excluded title keywords, location keywords, remote only, posted in the last N days.
• Include full job description: plain text, at the same price.
• Only output jobs not seen in earlier runs: monitoring mode, explained below.

How it finds the job board

1. If you give a job board URL, it uses that.
2. Otherwise it opens the company's /careers, /jobs and homepage, plus careers. and jobs. subdomains, and looks for links or embeds of a supported job board.
3. If that finds nothing, it tries the company name as the board ID, for example stripe.com -> greenhouse/stripe. A guessed Greenhouse board is only used when its job links or board name match the company.

Every row has matchedBy (url, careers-page or name-guess), so you can see how each board was found.

Output (one row per job)

company, title, department, location, country, remote, workplaceType, employmentType, salaryMin, salaryMax, salaryCurrency, salaryText, postedAt, updatedAt, url, applyUrl, ats, boardId, jobId, isNewSinceLastRun, and description when you turn it on.

Salary is filled only when the company publishes it on the board. That is mostly Greenhouse pay ranges and Ashby compensation, so mainly US roles.

Numbers from our test run (31 Oct 2026)

Input: 20 company domains, all jobs, no filters (airbnb.com, stripe.com, ramp.com, palantir.com, notion.so, gitlab.com, doctolib.fr, bunq.com, n26.com and others).

• 17 of 20 companies resolved. Revolut, Canva and Personio did not use a supported board. (An earlier build wrongly matched personio.com to a 1-job Recruitee test board; that guess is now rejected.)
• 2,749 open jobs in 2 min 40 s. Platform cost for the run: about $0.02.
• By job board: Greenhouse 1,699, Ashby 404, SmartRecruiters 314, Lever 301, Recruitee 26, Workable 6.
• 623 jobs (23%) had a salary range: 535 in USD, 25 EUR, 21 CAD, 16 GBP and a few others.
• 674 marked remote. 1,009 posted in October 2026.
• Stripe alone: 745 open jobs.

Monitoring: new and closed jobs

Turn on Only output jobs not seen in earlier runs and schedule the actor daily.

• The first run stores every job ID per company as a baseline.
• Each later run outputs only newly posted jobs.
• The SUMMARY record in the key-value store lists, per company, how many jobs are open, new and closed, with the titles of up to 100 closed jobs.

Use Monitor name to keep separate lists, for example one per client or per sales territory.

Pricing

$1.50 per 1,000 jobs saved. No monthly fee and no charge per run.

Jobs removed by your filters cost nothing, and so do already-seen jobs in monitoring mode. A daily monitor on 50 companies that post 20 new jobs a day in total costs about $0.90 a month.

Limits

• Only the seven job boards above. Companies on Workday, iCIMS, Taleo, SuccessFactors or their own system come back as "no supported job board found", at no charge.
• Workable rate-limits its public API. The actor waits and retries, so Workable companies are slower.
• A company that uses one of these boards for only part of its hiring will show only that part.

Data and terms

The actor uses the public job board endpoints that these systems provide so companies can show openings on their own websites. No login, no candidate data, and no personal data beyond what a company publishes in a job ad.

Feedback

Missing a company or a job board? Open an issue with the domain and we'll look at it.