Job Board Change Monitor — Greenhouse, Ashby & Lever

Keep job listings current with normalized vacancies and field-level change evidence from Greenhouse, Ashby and Lever. Sync a specialist job board, refresh a hiring-research database or route changes into an AI-agent workflow.

• Up to 10 boards in one batch; up to 5,000 jobs per board.
• Before/after values for titles, locations, descriptions, workplace type and structured compensation.
• Stable source IDs and content hashes for repeatable database upserts.
• Launch price: $0.01 per checked board, with all job rows included, plus $0.001 run start.
• Complete-scan protection preserves a healthy baseline during interrupted source checks.

First run

{
  "boards": [{"ats": "lever", "board": "spotify"}],
  "maxJobsPerBoard": 1000,
  "includeUnchanged": true
}

Use the company slug from its public career URL:

ATS	Public URL	Input
Greenhouse	job-boards.greenhouse.io/stripe	{"ats":"greenhouse","board":"stripe"}
Lever	jobs.lever.co/spotify	{"ats":"lever","board":"spotify"}
Lever EU	jobs.eu.lever.co/your-company	{"ats":"lever","board":"your-company","region":"eu"}
Ashby	jobs.ashbyhq.com/your-company	{"ats":"ashby","board":"your-company"}

Get jobs and change evidence

Each dataset row includes id, ats, board, title, location, departments, employmentType, workplaceType, compensation, descriptionText, jobUrl, applyUrl, contentHash, warnings and observedAt.

The first run uses change: current. With a previous snapshot, rows are added, changed, unchanged or removed. Changed rows include field names and before/after values:

{
  "id": "greenhouse:global:example:123",
  "change": "changed",
  "changedFields": ["title"],
  "fieldChanges": {"title": {"before": "Engineer", "after": "Senior Engineer"}},
  "sourceComplete": true
}

contentHash covers normalized comparison fields. Plain-text conversion and whitespace normalization filter formatting-only description changes. Native structured workplace and compensation values are preserved; unavailable values use null.

SUMMARY contains board completeness, counts, errors, usage and budget outcomes. SNAPSHOT is issued when every requested board completes within size and spending caps. Export JSON, JSONL, CSV or Excel; use JSON for nested change evidence.

Automate repeatable synchronization

import os
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])
actor = client.actor("gertner-data/job-board-changes")
inputs = {"boards": [{"ats": "lever", "board": "spotify"}]}
first = actor.call(run_input=inputs)
store = client.key_value_store(first["defaultKeyValueStoreId"])
summary = store.get_record("SUMMARY")["value"]
if not summary["snapshotAvailable"]:
    raise RuntimeError("Inspect SUMMARY before advancing the baseline")
baseline = store.get_record("SNAPSHOT")["value"]
second = actor.call(run_input={
    **inputs, "previousSnapshot": baseline, "includeUnchanged": False
})
for job in client.dataset(second["defaultDatasetId"]).iterate_items():
    print(job["change"], job["id"], job["changedFields"])

Your backend, n8n or Make owns the baseline: schedule → load snapshot → run Actor → route changes → save the next complete snapshot. Keep the previous healthy snapshot during failures. Supply the same board list between comparisons; omit previousSnapshot to reset.

Pricing

Launch price: $0.001 run start + $0.01 per checked board, platform usage included. All job rows and snapshot data for each checked board are included.

Boards	Event price
1	$0.011
3	$0.031
10	$0.101

A valid empty board or partially read board with usable jobs counts as checked. A board without usable source data receives no board event. The start event applies independently. includeUnchanged controls output, not the board price. Set the maximum run charge to limit spending before the next board is fetched.

API coverage and run settings

Reads public listed postings from the three ATS posting APIs, including Lever EU. Removal events mean a posting left a completely read source board. Partial boards retain structured diagnostics; complete snapshots keep subsequent comparisons reproducible.

1–10 boards, 5,000 jobs per board, 15 MB per response, 20 MB downloaded per board, 6 MB combined baseline, 100 KB job text and 12 MB comparison output per board. Split large boards into appropriate runs and inspect SUMMARY for capped results. Lever uses live pagination; repeat a complete observation for consequential removal decisions.

20-second requests, two bounded retries for network errors/429/5xx. Default 512 MB; recommended timeout 600 seconds. Logs exclude job descriptions and query secrets. Inputs and results use your Apify retention settings. Use permitted public sources and their data-use terms.

For http_error 404, check the board slug or ATS migration. For max_jobs_reached, adjust the job limit within run caps. For snapshotAvailable: false, retain the previous complete baseline. Report issues with the run ID and SUMMARY error codes.

Support: Evgeny Gertner.