Unified ATS Job Feed avatar

Unified ATS Job Feed

Pricing

from $30.00 / 1,000 results

Go to Apify Store
Unified ATS Job Feed

Unified ATS Job Feed

Combines public Greenhouse, Lever, and Ashby job board JSON APIs into a normalized job dataset with first-seen tracking, per-board failure reporting, and run summaries.

Pricing

from $30.00 / 1,000 results

Rating

0.0

(0)

Developer

Jason Ramirez

Jason Ramirez

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

2 days ago

Last modified

Categories

Features

Limited permissions

Share

Track company job postings across Greenhouse, Lever and Ashby in one feed. Each run returns normalized job rows with first-seen dates and new-posting flags, plus a run summary counting new and closed postings per board, so you can watch hiring changes at the companies you track without checking three board systems by hand.

An Apify Actor for lead gen and recruiting-tools builders who need cross-ATS job rows. It reads public JSON job boards from Greenhouse, Lever, and Ashby and emits one normalized dataset item per job. It does not scrape HTML pages or require source credentials.

Input

{
"companies": [
{ "platform": "greenhouse", "board": "example-greenhouse-board" },
{ "platform": "lever", "board": "example-lever-site" },
{ "platform": "ashby", "board": "example-ashby-board" }
],
"maxJobsPerCompany": 500,
"includeDescription": false
}

Replace placeholders with actual public board tokens. Companies is required and nonempty. Exact duplicate platform/board pairs are fetched once; whitespace around board tokens is trimmed. maxJobsPerCompany must be a positive integer and defaults to 500. It caps rows emitted per board, after sorting. Full public API responses still must be downloaded and normalized for accurate closure counts. It is not a network or memory cap.

Output schema

FieldTypeMeaning
platformstringgreenhouse, lever, or ashby
companystringInput board token, not a resolved legal company name
jobIdstringSource-native posting ID
titlestringPosting title
urlstringAbsolute apply or job URL
locationstringBest available location, empty when missing
locationTypestringremote, onsite, hybrid, or unknown
departmentstring or nullFirst or source department
teamstring or nullSource team; Greenhouse uses first office
employmentTypestring or nullSource value when available
postedAtISO string or nullSource publication/creation date when available
updatedAtISO string or nullSource update date when available
descriptionTextstringBasic HTML-to-text conversion; not a full HTML parser
descriptionOmittedbooleanTrue when full descriptions were not requested
firstSeenAtISO stringFirst observation in the retained snapshot
lastSeenAtISO stringCurrent run timestamp
isNewbooleanPosting absent from the previous snapshot

With includeDescription false, descriptionText is at most 500 JavaScript string characters and descriptionOmitted is true, even for short descriptions. With true, the full converted text is emitted. Missing optional fields do not fail a board. Missing IDs, titles, valid absolute URLs, or expected top-level arrays fail that board to avoid false closures. No closedSinceLastRun field is emitted on job items.

Rows sort by platform, company, postedAt descending (null last), then jobId lexically. Adapter modules expose endpoint(board) and parse(payload); new platforms can map to the same normalized fields without changing the dataset schema. Workable, SmartRecruiters, and Recruitee are not implemented.

Velocity and honest receipts

The default key-value store retains SNAPSHOT-{platform}-{board}, containing runAt and a jobs map of jobId to firstSeenAt/title. New IDs receive the current timestamp and isNew true; existing IDs retain firstSeenAt. IDs absent from a successfully fetched full board count as closed. A successful empty board closes all previous IDs. Failed boards leave their snapshots untouched and contribute no closures.

The full fetched board is snapshotted even when output is capped. Therefore a previously observed job can first appear in the emitted dataset with isNew false. newJobs counts new emitted rows; closedJobs counts closures across complete successful boards. Reusing the same persistent default store is required for history across runs. A new default store starts fresh history; this is not a global history service.

A single summary per completed run goes to the named run-summary dataset and LAST_RUN_SUMMARY in the default store. Fields: runAt, boardsRequested (unique input boards), boardsSucceeded, boardsFailed (array of platform/company/error), totalJobs (emitted rows), newJobs, closedJobs. A nonexistent board fails that board and appears in the summary. HTTP, JSON, shape, and timeout errors are isolated per board. Each request times out after 30 seconds; there are no retries. Output or persistence failures abort the run and may prevent a final summary. Dataset writes and snapshots are not transactional; retries after partial storage failure may repeat rows.

Honest limits and publication

Public boards only. Endpoint shapes in BRIEF.md are ASSERTED, not live-verified in this build. Test fixtures are invented realistic payloads, not recordings of requests. Unknown location types become unknown; dates are not inferred from observation time. Descriptions may contain text published by the source; no personal records are requested. Basic entity conversion does not decode every HTML entity. Greenhouse employment type and posted date can be absent.

Intended pricing when published: pay per job row emitted. Store pricing, billing event configuration, and publication are owner gates; this local build does not charge money or configure paid events. No demand or revenue claims are made.

Local run

Use Node.js >=18. With dependency access available, run npm install, then npm test. Tests use node:test, local synthetic fixtures, and injected HTTP responses; they require neither Apify installation nor network access. Run npm start after installing dependencies and supplying an Apify local INPUT.json in the default key-value store (normally storage/key_value_stores/default/INPUT.json). The Actor uses the Apify SDK v3 and default dataset/store. The Dockerfile uses the Apify Node 22 base image.

The inherited frontend directories and configs are outside this Actor's entrypoint and test command. The inherited package-lock.json was removed because it described a different application. Generate a fresh lockfile after dependency installation. An actual SDK startup, container build, live requests, and billing setup are not verified by the offline tests.