Company Jobs Feed: Greenhouse, Lever & Ashby

Live job postings taken straight from companies' own applicant tracking systems, not scraped from a job board.

Most job scrapers hit Indeed, LinkedIn or Glassdoor. That means anti-bot defences, proxies, breakage, and terms nobody enjoys reading. Companies already publish their vacancies through ATS platforms with official public APIs meant for aggregators. This reads those instead.

The practical result: no proxies, no blocking, no stale cache of a scraped page. The data is what the company itself published, in the format it published it.

What it covers

• Greenhouse, Lever and Ashby
• Tracked companies including Stripe, OpenAI, Databricks, Anthropic, Datadog, MongoDB, Notion, Snowflake, GitLab, Airbnb, Coinbase, Pinterest, Elastic, Plaid, Deliveroo and more
• Roughly 8,000 open postings at any time, about a third of them remote

Input

{
  "titleKeywords": ["engineer", "designer"],
  "locationKeywords": ["london", "remote"],
  "remoteOnly": true,
  "postedWithinDays": 7,
  "maxPerCompany": 200
}

Leave companies empty and it searches every tracked company at once. That is usually what you want.

To target specific employers, pass them as ats:slug:

{ "companies": ["greenhouse:stripe", "ashby:openai", "lever:spotify"] }

• titleKeywords, locationKeywords: case-insensitive, match any
• remoteOnly: keep only postings the company flags as remote
• postedWithinDays: 0 means no recency filter
• includeDescription: adds full description text. Much larger output
• maxPerCompany: cap per employer

Output

{
  "company": "Notion",
  "companySlug": "notion",
  "ats": "ashby",
  "jobId": "8fb1615c-34bf-47c4-a1d1-b7b2f836bbd3",
  "title": "Software Engineer, Developer Platform",
  "department": "Engineering",
  "location": "San Francisco, California",
  "isRemote": false,
  "employmentType": "FullTime",
  "url": "https://jobs.ashbyhq.com/notion/8fb1615c-...",
  "applyUrl": "https://jobs.ashbyhq.com/notion/8fb1615c-.../application",
  "publishedAt": "2026-08-19T14:02:07Z",
  "checkedAt": "2026-08-24T19:40:00Z"
}

All three systems are normalised to one shape, so you never write per-provider parsing.

isRemote can be absent. Not every ATS states it, and guessing would be worse than saying nothing. Absent means unknown, not false.

Good for

• Job boards and aggregators: a clean feed that does not break when a site changes markup
• Recruiting and market research: who is hiring, for what, where, and since when
• Tracking a company or a competitor: pass their slug and schedule it
• Feeding an AI agent: one consistent schema across employers

Scheduling

Schedule it and diff the results. publishedAt and checkedAt on every row make it easy to spot what is new since the last run.

Limitations, stated plainly

• Coverage is the tracked company list, not the whole job market. If an employer does not use Greenhouse, Lever or Ashby, or is not yet in the list, it will not appear. The list grows.
• Locations are free text, exactly as the employer wrote them. "London", "London, UK" and "UK Remote" are all possible and are not normalised.
• No deduplication across companies. The same role posted by an agency and by the employer appears twice.
• No salary data. Most of these feeds do not carry it.
• A posting appearing here means the company published it. It does not guarantee the role is still open at the moment you read it.

Disclaimer

This is an unofficial tool. It is not affiliated with, endorsed by, or sponsored by Greenhouse Software, Lever, Ashby, or any of the employers whose postings it returns. It reads their public job board APIs. Job adverts are company-published content and the output contains no personal data.