Lever Jobs Scraper & API - Open Roles, Any Company

For recruiters, sales teams selling into hiring companies, and anyone building a jobs feed: every open role on any company's Lever job board, read straight from Lever's own public board API — title, URL, department, employment type, workplace type and posted date, one identical row shape for every board.

Measured, not promised. A run on 2026-09-01 with the shipped build and the default input returned 155 job rows across 2 boards. title, jobUrl, postedAt, department and workplaceType came back on 100% of rows from both; employmentType on 97% and 96%. Up to 200 boards per run.

And what Lever does not publish is said out loud: isRemote is set only from Lever's own workplaceType, so it depends on how each company fills in its board — 34% on one default board and 0% on the other. Lever publishes no structured salary at all, so salaryMin and salaryMax stay null rather than being guessed from the advert text.

No HTML scraping, no proxies, no API keys, nothing that breaks when a careers page is redesigned. Export as JSON, CSV or Excel.

Unofficial. This Actor is not affiliated with, endorsed by, or sponsored by Lever. All product names are trademarks of their respective owners.

Don't know a company's board name? ATS Job Board Finder turns a domain into the board name this Actor takes.

---

It never returns an empty result to mean two different things

Most job-board Actors return an empty dataset when something goes wrong, and an empty dataset when there is genuinely nothing to report. Those are not the same answer, and if you are monitoring hiring activity the difference is the whole point.

Every board you ask for produces at least one row, tagged in the rowType column:

rowType	ok	What it means	Charged?
job	true	A real open job	Yes
no-open-roles	true	The board is live and has no open jobs (or none matching your filters)	No
board-not-found	false	No board with that name exists on Lever	No
unavailable	false	Lever could not be reached, or the entry could not be read as a board; reason says why	No
budget-reached	false	The run hit the Max total charge you set. Names every board that was cut short and every board not fetched because of it	No

You are charged only for job rows.

---

What Lever actually publishes - measured, not estimated

Run on 2026-09-01 with the shipped build and the default input: 155 job rows across 2 boards.

	spotify (88)	matchgroup (67)
title, jobUrl, postedAt, department	100%	100%
employmentType	97%	96%
workplaceType	100%	100%
isRemote	34%	0%
salaryMin / salaryMax	0%	0%

isRemote is set only from Lever's own workplaceType, so it depends on how each company fills its board - 34% on one of the two default boards and 0% on the other. Lever publishes no structured salary, so salaryMin stays null everywhere. Lever returns createdAt as a UNIX millisecond number; here it is ISO 8601 UTC like every other source.

---

Input

Field	Type	Default	What it does
boards	array	lever:spotify, lever:matchgroup	One entry per company board, up to 200
titleContains	string	(empty)	Keep only jobs whose title contains this text (case-insensitive)
departmentContains	string	(empty)	Keep only jobs whose department or team contains this text
locationContains	string	(empty)	Matches the primary location and every secondary location
remoteOnly	boolean	false	Keep only jobs Lever itself flags as remote - never a guess
postedWithinDays	integer	0 (off)	Keep only jobs first published within N days. Unknown dates are excluded rather than assumed recent
maxJobsPerBoard	integer	0 (no limit)	Cap of job rows per board, applied after filtering
includeDescription	boolean	false	Add the full description as plain text in descriptionText
maxRetries	integer	3	Total attempts per board for temporary errors (HTTP 429 / 5xx). A 404 is a permanent answer and is not retried

Naming a board: lever:spotify, or paste https://jobs.lever.co/spotify, or just spotify

---

Output columns

rowType, ok, reason, ats, companySlug, companyName, jobId, title, department, team, employmentType, locationRaw, locationsAll, country, city, workplaceType, isRemote, remoteHint, salaryMin, salaryMax, salaryCurrency, salaryInterval, salaryText, postedAt, updatedAt, jobUrl, applyUrl, descriptionText, source

All 29 columns are present on every row, in the same order - including the marker rows. The Greenhouse, Lever, Ashby, Workable and Workday Actors all return these same 29 columns, so adding another ATS never means rewriting your parser.

Two things it refuses to invent

• isRemote is set only when Lever itself says so. A guess based on the location text lives in a separate column called remoteHint, so you always know which is which. A job located in "Remote, Oregon" is not a remote job.
• salaryMin / salaryMax come only from structured compensation fields. Nothing is parsed out of prose, and a missing figure stays null - never 0.

---

How it is checked

• 148 automated checks across 3 files, run before every publish
• 50 deliberate defects injected into the source; all 50 were caught by those checks (re-run and verified 2026-09-03)
• Only public job boards are read. Nothing needs a login, and no private ATS data is accessed
• Every request has a 45-second timeout, so a silent server cannot hang the run
• Up to 200 boards per run

---

The rest of the family

Same 29 columns, same row rules, one ATS each:

• greenhouse-jobs - Greenhouse boards
• ashby-jobs - Ashby boards
• workable-jobs - Workable boards
• workday-jobs - Workday boards
• ats-jobs-api - all five in one run, for a mixed list
• ats-board-finder - a company domain to the ATS and board name it uses

Support

Found a board that behaves oddly? Open an issue on the Issues tab and include the board name - it will be reproduced against the live API before anything is changed.