If you are building a job aggregator, a hiring tracker or a "who is hiring" alert, you do not need to scrape careers pages. Most companies host their openings on an applicant tracking system (ATS), and six of the big ones serve every company's jobs as plain JSON: no key, no login, no HTML. The endpoints exist so that companies can embed their own openings on their own sites.
The catch has always been the other half of the problem: which companies are on which system, and under what identifier? This post covers both. The endpoints are below, and the list of companies is open data: https://github.com/storesignalhq/ats-job-boards (CSV, CC BY 4.0).
The six endpoints
| System | Endpoint | Good to know |
|---|---|---|
| Greenhouse | https://boards-api.greenhouse.io/v1/boards/{slug}/jobs | ?content=true adds descriptions |
| Lever | https://api.lever.co/v0/postings/{slug}?mode=json | EU accounts are on api.eu.lever.co |
| Ashby | https://api.ashbyhq.com/posting-api/job-board/{slug} | ?includeCompensation=true adds pay ranges |
| Workable | https://apply.workable.com/api/v1/widget/accounts/{slug} | ?details=true adds descriptions |
| SmartRecruiters | https://api.smartrecruiters.com/v1/companies/{slug}/postings | paginated, 100 per request |
| Recruitee | https://{slug}.recruitee.com/api/offers/ | descriptions included |
Try one in a terminal:
curl -s https://boards-api.greenhouse.io/v1/boards/stripe/jobs | python -m json.tool | head -30 {slug} is the company's identifier inside that system. It is usually the company name (stripe, openai, spotify), but not always, and there is no official index of them. That is what the list is for.
The list
boards.csv has one row per company board: system, slug, company name, number of open jobs when the board was checked, the human-readable board URL and the JSON endpoint.
ats,slug,company_name,open_jobs,board_url,api_url greenhouse,stripe,Stripe,... What is in it, as of 16 November 2026:
| System | Boards | Open jobs |
|---|---|---|
| Greenhouse | 5,347 | 191,722 |
| Ashby | 3,322 | 62,586 |
| Workable | 3,177 | 132,874 |
| Recruitee | 2,041 | 33,679 |
| SmartRecruiters | 1,461 | 244,949 |
| Lever | 299 | 27,248 |
| Total | 15,647 | 693,058 |
Two things the numbers show. First, the typical board is small: the median company has 9 open jobs, and 1,458 boards have exactly one. Only 1,104 boards have 100 or more. Second, the systems are not alike: SmartRecruiters has fewer companies than Greenhouse but more jobs, because it is used by large employers with thousands of store and warehouse roles.
How it was built
- Candidate URLs come from the Common Crawl URL index: every crawled URL on the six systems' board hosts, reduced to unique slugs.
- Each candidate was requested once from the public endpoint, slowly.
- Only boards that answered with at least one open job were kept.
It is a snapshot, so it ages. Companies switch systems and close boards; expect a few percent of rows to return 404 or an empty list a month later.
Read jobs for any board in 30 lines
The repository includes fetch_jobs.py, standard library only. The core of it is one small function per system, because every system names its fields differently:
def greenhouse(slug): for j in get(f"https://boards-api.greenhouse.io/v1/boards/{slug}/jobs")["jobs"]: yield j["title"], (j.get("location") or {}).get("name"), j["absolute_url"], j.get("first_published") def ashby(slug): for j in get(f"https://api.ashbyhq.com/posting-api/job-board/{slug}")["jobs"]: yield j["title"], j.get("location"), j["jobUrl"], j.get("publishedAt") def lever(slug): for j in get(f"https://api.lever.co/v0/postings/{slug}?mode=json"): yield j["text"], (j.get("categories") or {}).get("location"), j["hostedUrl"], j.get("createdAt") Usage:
python fetch_jobs.py greenhouse:stripe ashby:openai lever:spotify python fetch_jobs.py --ats greenhouse --top 50 --title "data engineer" > jobs.jsonl A run made on 19 November 2026 returned 780 jobs for Stripe, 834 for OpenAI and 90 for Spotify.
Things that will bite you at scale
Reading three boards is trivial. Reading 15,000 is where the differences between systems start to matter:
- Rate limits are per system, and uneven. Workable answers with HTTP 429 after short bursts, and the block can last a day. Greenhouse and Ashby tolerate far more. Keep a separate, slow queue per system and stop when you see 429 instead of retrying.
- SmartRecruiters cannot tell you a company does not exist. An unknown slug returns an empty list with HTTP 200, the same as a company with no openings.
- Pagination only exists on SmartRecruiters, and descriptions there cost one more request per job.
- Dates mean different things. Lever gives a creation timestamp in milliseconds, Workable a date without time, Greenhouse both a first-published and an updated time.
- Remote is not always a field. Greenhouse has no remote flag in the list response, so there you are matching the word in a location string.
If you would rather not run the crawl yourself
The same directory is built into a hosted scraper, Career Site Jobs Search on Apify. It reads all boards (or your own list) live, normalizes the six formats into one schema and filters by title keywords, location, remote and posting date. It costs $1 per 1,000 jobs returned, plus a flat $0.30 for a run over more than 500 boards. A full scan of the directory takes about 45 minutes. A worked example, with a script that groups matching jobs by company: Who is hiring for RevOps right now?
The list itself stays free. If your company's board should not be in it, open an issue in the repository and it will be removed.