Career Site & ATS Jobs Scraper - Greenhouse, Lever, Ashby and more

Give this Actor a company career page, a public ATS board URL or just a domain. It reads the board the company already publishes and returns every open role, one row per job, with location, department, employment type, salary, a description excerpt and the skills, benefits and experience requirements we read out of the posting.

Point it at the same boards again next week with the previous dataset id and it also tells you what changed: which roles are new, which look like reposts, which disappeared, and how each company's open headcount moved.

Supported public boards: Greenhouse, Lever, Ashby, SmartRecruiters, Personio, Recruitee, Workable.

Unofficial. This Actor is not affiliated with, endorsed by or sponsored by Greenhouse, Lever, Ashby, SmartRecruiters, Personio, Recruitee, Workable or any employer named in your input. All product names are trademarks of their respective owners and are used only to describe which public boards are read.

What it is good for

• Sales and recruiting signals - a company that opened twelve engineering roles this month is a different prospect than one that closed five.
• Job boards and aggregators - keep a live copy of hundreds of company boards without writing seven different clients.
• Market research - salary, seniority, department and location across a whole portfolio of companies.
• Competitor hiring watch - run it weekly, pass the previous dataset id, and read only the diff.

Input - every field is an array

Field	What it does
boardUrls	Public board URLs, or platform:token shorthand such as greenhouse:stripe. Up to 50 per run.
companyDomains	Domains such as acme.com. Each domain is tried against every supported board pattern, so one domain can return a whole board.
atsPlatforms	Restrict to greenhouse, lever, ashby, smartrecruiters, personio, recruitee, workable. Empty means all.
titleIncludes / titleExcludes	Keep or drop jobs by title keyword.
locationIncludes	Keep jobs whose location contains one of these terms.
postedWithinDays	One integer. 0 disables the date filter.
maxResults	One integer, 1 to 5000. Upper bound on charged rows.
includeDescription	One boolean. Description excerpt and the derived skill, benefit, language and experience fields.
fetchDetailPages	One boolean. Fetch the public job page for boards whose list feed carries no description (SmartRecruiters, Personio).
previousDatasetIds	One dataset id from an earlier run of this Actor. This is what turns on change detection.
comparisonModes	off or compare. Use compare together with previousDatasetIds.
detectClosures	One boolean. Emit closure_suspected rows. Those rows are free.
proxyModes	NONE, DATACENTER or RESIDENTIAL. Public ATS endpoints answer a direct connection, so NONE is the default.

Accepted board URL shapes include https://job-boards.greenhouse.io/<token>, https://boards-api.greenhouse.io/v1/boards/<token>/jobs, https://jobs.lever.co/<token>, https://jobs.ashbyhq.com/<token>, https://jobs.smartrecruiters.com/<Company>, https://<company>.jobs.personio.de/, https://<company>.recruitee.com/ and https://apply.workable.com/<token>/.

Example

{
  "boardUrls": ["https://job-boards.greenhouse.io/stripe", "https://jobs.lever.co/leverdemo"],
  "companyDomains": ["vanta.com"],
  "titleIncludes": ["engineer", "data"],
  "postedWithinDays": [30],
  "maxResults": [500],
  "includeDescription": [true],
  "previousDatasetIds": ["PASTE_YOUR_PREVIOUS_DATASET_ID"],
  "comparisonModes": ["compare"]
}

Output

One row per job. A few of the fields:

Field	Meaning
jobKey, jobId, url, applyUrl	Stable key and the public links. jobKey is platform:token:jobId.
title, companyName, location, isRemote, workModel	The basics, normalised across all seven platforms.
department, team, jobFunction, employmentType, seniorityLevel	Classification as the board publishes it.
salary, salaryMin, salaryMax, salaryCurrency, salaryInterval, hasSalary	Compensation when the board exposes it.
descriptionSnippet, descriptionChars, descriptionTruncated, descriptionUrl	First 500 characters of the posting, its true length, and the public page that holds the rest.
skillsMentioned, benefitsMentioned, languagesMentioned, yearsOfExperience, requiresVisaSponsorshipMention	Facts we derive from the posting text so you can filter without the prose.
postedAt, updatedAt, validThrough, ageDays, observedAt	Dates, always ISO 8601.
atsPlatform, companyBoardToken, boardUrl, sourceDomain, evidenceUrls	Exactly which public page each field came from.
isNew, isReposted, suspectedClosed, previousJobId, postedAtChanged, changeReasons	Change detection.
companyCurrentCount, companyPreviousCount, companyDelta	Open roles per company, now and in your baseline.
comparisonStatus, coverageComplete, scopeFingerprint	Whether the comparison is trustworthy and why.
fieldAvailability	Per field: observed, not_measured or excluded_by_policy.

Aliases such as jobTitle, company, organization, jobUrl, link and postedDate are included so the rows drop straight into pipelines built for other job scrapers.

Why there is no full description field

The posting text is the employer's copyrighted work. We do not republish it. Each row carries a 500-character excerpt, the real character count, the structured facts we derive locally, and descriptionUrl, the employer's own public page where the full text lives. descriptionText, description and descriptionHtml exist in the schema but are always null with excluded_by_policy, so a pipeline expecting those keys will not break and will not silently receive republished prose.

Change detection is honest about what it knows

• No previousDatasetIds - every change field is null and comparisonStatus is no_baseline. Nothing is guessed.
• Baseline scope differs from this run - scope_mismatch, change fields stay null.
• This run or the baseline did not cover the whole board - incomplete, change fields stay null and no closure rows are produced.
• comparable - isNew, isReposted, companyDelta and closure rows are filled in.

suspectedClosed means the job was in your baseline and is absent from a board listing that was read in full in this run. It is a strong signal, not a statement from the employer, which is why the field is named suspected.

The run summary in the key-value store record OUTPUT reports per board how many jobs were discovered, which boards failed and why, and how many descriptions were fetched. It keeps two coverage flags apart: coverageComplete says the board listings were read in full, which is what change detection and closure rows depend on, while descriptionsComplete says every description we went looking for came back. A board that answered its listing but rate-limited one detail page leaves coverageComplete: true and descriptionsComplete: false.

Pricing

Event	What triggers it
Actor start	Charged once per run by the platform, based on memory.
Job result	One validated, unique open job row delivered to the dataset.

You are not charged for closure rows, duplicate rows, boards that returned an error, boards that do not exist, or runs that produce no jobs beyond the start event. Tiered pricing applies, so paid Apify plans pay less per row than the free tier. Current prices are on the Actor's pricing tab and always take precedence over this page.

If you set a maximum total charge for the run, the Actor plans its output to fit inside it, stops cleanly, keeps what it already delivered and marks coverageComplete: false.

How it collects

Only public, unauthenticated endpoints that each ATS publishes for its own job boards, with an honest bot user agent, at most a few requests per second, one request at a time per board and exponential backoff on transient errors. There is no login, no cookie, no captcha solving, no account pool and no IP rotation to get around a block. A board that refuses us is reported as refused.

Personal data is out of scope by design: contact emails, phone numbers and recruiter identities are never collected, and those fields are always null with excluded_by_policy in fieldAvailability.

Your responsibility as the operator

You choose the companies and boards. You are responsible for having the right to collect and use that data, for complying with the terms of the sites you point this Actor at, and for any onward use of the output, including GDPR and equivalent obligations where you are the data controller. We provide the tool and report what each public page returned.

Leave a review

A short review on the Reviews tab, good or bad, helps other people find this Actor and tells us what to build next. Found a bug or need a feature? Open an issue on the Issues tab and we'll look into it.