An open dataset and Python toolkit for job data from ATS platforms and public sources.

ats-scrapers provides two layers:

• A free, hosted dataset with 4.2M+ live jobs from 63,000+ companies across 49 sources.
• More than 50 reusable scraper adapters, including Workday, Greenhouse, Lever, Ashby, SmartRecruiters, and SuccessFactors.

Jobs are collected from ATS endpoints, company career sites, and public job feeds, then normalized into one typed schema. Querying the hosted dataset requires no API key or account.

pip install ats-scrapers

The package is imported as ats_scrapers. Optional extras add only what you need:

pip install "ats-scrapers[parquet]"   # query the full Parquet snapshot
pip install "ats-scrapers[scrapers]"  # run the scraper library
pip install "ats-scrapers[all]"       # install every runtime extra

from ats_scrapers import search

# Per-source searches work with the base install.
jobs = search(
    query="machine learning engineer",
    location="Paris",
    ats="greenhouse",
    limit=100,
)

# The result is a pandas DataFrame.
print(jobs[["company", "title", "location", "apply_url"]])

For practical full-dataset queries, install the parquet extra. The base install is intended for smaller per-source CSV slices.

from ats_scrapers import search

jobs = search(query="data engineer", remote=True, salary_min=80_000)

The live manifest contains current row counts and artifact URLs. See the job schema for field definitions and normalization rules.

You don't need to know which ATS a company uses. Paste its careers URL:

from ats_scrapers import get_scraper_for_url

scraper = get_scraper_for_url("https://jobs.ashbyhq.com/openai")
jobs = scraper.fetch()

Or look it up by name in the hosted companies directory (63,000+ tenants):

from ats_scrapers import find_company
from ats_scrapers.scrapers import get_scraper

find_company("openai")          # → ats="ashby", slug="openai", url=...
scraper = get_scraper("ashby", "openai")
jobs = scraper.fetch()

Scraper classes are also available directly:

from ats_scrapers.scrapers import GreenhouseScraper

jobs = GreenhouseScraper("anthropic").fetch()

Scrapers are async-first — in async code (or for concurrency) use await scraper.afetch() instead. The sync fetch() also works from inside a running event loop (Jupyter, FastAPI): it transparently runs on a worker thread.

Scraper adapters include:

• Major ATS platforms: Greenhouse, Lever, Ashby, Workday, SmartRecruiters, SuccessFactors, Oracle, iCIMS, Workable, Personio, and more.
• First-party company APIs: Amazon, Apple, Google, TikTok, and Uber.
• Public and regional sources: EURES, Bundesagentur, Arbetsformedlingen, Welcome to the Jungle, and others.

Run python -c "from ats_scrapers import list_ats; print(*list_ats())" for the sources currently present in the hosted dataset. The dataset can list a source before this package ships a scraper for it — search() still returns those rows; only building your own scraper needs one. To skip sources without a scraper:

from ats_scrapers import list_ats
from ats_scrapers.scrapers import ScraperRegistry, get_scraper

scrapeable = [a for a in list_ats() if ScraperRegistry.has_scraper(a)]

Contributions can add a source, improve an existing scraper, or add companies to the CSV inventories in ats-companies/. The scraper API is intentionally tiny: subclass BaseScraper, set ats, and implement async def afetch() using self.make_fetcher() for HTTP — retries, backoff, and error mapping come for free. See src/ats_scrapers/scrapers/greenhouse.py for a compact reference and the Job model in src/ats_scrapers/models.py for the schema you populate.

git clone https://github.com/kalil0321/ats-scrapers
cd ats-scrapers
uv sync --extra dev
uv pip install -r pipeline/requirements.txt  # repo-only ops/publisher tests
uv run pytest
uv run ruff check .

MIT