WARN Layoff Notices Scraper — US Layoff & Plant Closure Data from Official State Sources

Know which companies are about to lay people off, 60–90 days before it happens, across 10 states in one run. A WARN notice is a statutory filing: the employer, the site, how many people, and the date it takes effect. Recruiters use it to reach experienced candidates before they hit the market; sales teams use it as the strongest public distress signal on an account.

Data comes from 10 official state government sources, normalized into one schema. No login, no API key, no HTML scraping of news sites.

A real run, with the real number

Every WARN notice filed in the covered states in the last 90 days:

{
    "noticeDateFrom": "2026-05-16",
    "maxResults": 20000
}

Result on 2026-08-14: 400 notices, from 10 of 10 state sources, cost $2.00. That is the whole national picture for a quarter, for the price of a coffee — and the run reports which states succeeded so you always know what you are looking at.

See real records for 13 cents before you commit

Set maxResults to 25 and run it once. You are charged $0.005 per notice — so a 25-record look costs $0.125, and a run that matches nothing costs only Apify's fraction-of-a-cent start event.

What one record actually looks like

Real output from a live run, trimmed to the fields most people check first:

{
    "state": "CA",
    "company": "McDonald's Corporation",
    "city": "City of Industry",
    "county": "Los Angeles",
    "employeesAffected": 2,
    "noticeDate": "2026-06-30",
    "layoffDate": "2026-09-01",
    "closureType": "Closure",
    "industry": "72 Accommodation and Food Services",
    "sourceUrl": "https://edd.ca.gov/en/jobs_and_training/Layoff_Services_WARN/"
}

Every record carries both the normalized value (ISO dates, integer counts, Closure/Layoff classification) and the raw value exactly as the state published it, so nothing is lost to normalization. Every record also names the official source it came from.

What a real job costs

Pricing is $5.00 per 1,000 notices ($0.005 each), charged only on notices actually returned:

What you pull	Notices	Cost
A look at the output before buying	25	$0.13
All 10 states, last 90 days	400 (measured)	$2.00
A year of one state	varies	$0.005 each

There is no subscription — just the per-notice price plus Apify's fraction-of-a-cent platform events (an Actor start event and $0.00001 per result, about 0.2% on top). A weekly monitoring schedule that finds 30 new notices costs about 15 cents.

On very large pulls: Apify caps each run with a maximum-charge limit. If a run reaches it, the run stops early and returns fewer records than your maxResults — the run's status message will tell you when this happens. For big pulls, raise the maximum-charge setting on the run (or your plan limit) before starting it.

Who uses WARN data?

• Recruiters & staffing agencies — a WARN notice is a public list of experienced people who will be job-hunting on a known date, 60-90 days before they hit the market. Filter by state, industry, and size to find fresh talent pools first.
• B2B sales & account teams — a WARN filing is one of the strongest publicly available distress/change signals for an account: churn risk for vendors, an opening for cost-cutting products, facility consolidation intelligence.
• Journalists & researchers — track layoff waves by state, industry, and month from primary government sources, not press releases.
• Investors & analysts — plant closures and mass layoffs, straight from the statutory filings, weeks before they show up in quarterly numbers.
• Local economic development & workforce boards — monitor incoming dislocations in your region.

What data do you get?

One result per WARN notice (per affected site, where the state files them that way):

Field	Example
state / stateName	TX / Texas
company	Essendant
address / city / county	300 Central Ave. / Irving / Dallas
employeesAffected	136
noticeDate	2026-08-03 (ISO; see coverage notes for month-only states)
noticeMonth / noticeDateRaw	2026-08 / date exactly as the state published it
layoffDate / layoffDateRaw	2026-10-03 / 10-03-26 thru 10-13-26
closureType / closureTypeRaw	Closure / Closure Permanent
industry	Health Care and Social Assistance (where published)
source / sourceUrl	The official state source each record came from

State coverage — honest and specific

There is no federal WARN database. Every state publishes its own portal in its own format — Excel workbooks, CSV exports, Google Sheets, paginated HTML, PDF-only filings. This Actor wraps the machine-readable ones, normalizes the fields, and tells you exactly which states it can and cannot cover.

Covered now (10 states, ~45% of US employment):

State	Official source	Format	History available
California	EDD WARN Report	Excel, updated Tue & Thu	Current fiscal year only (July 1 onward) — EDD publishes prior years as PDF
Texas	TWC WARN Act Listings	Excel per year	Multi-year (per-year files)
Florida	FloridaCommerce REACT	HTML per year	Multi-year (verified to 2015)
Illinois	DCEO monthly WARN reports	Excel per month	Recent years (older months are PDF-only)
Pennsylvania	L&I WARN notices page	HTML	Current + previous year; notice date = posting month
New Jersey	NJ DOL WARN Archive	Excel workbook	Since 2004; notice date = posting month
Virginia	Virginia Works	CSV export	Multi-year
Wisconsin	DWD WARN notices	Official data feed	Since 2020
Colorado	CDLE layoff/WARN list	Google Sheets per year	~3 years
Maryland	DWDAL WARN log	HTML per year	Multi-year

Not covered yet — and why (verified individually, August 2026):

• New York — NY DOL moved WARN data into a Tableau dashboard behind bot protection; the underlying filings are per-notice PDFs. On the roadmap.
• Ohio — JFS reorganized its site in 2026; the public notice listing URLs currently return 404. Will be added when the listing resurfaces.
• Georgia — TCSG publishes notices through a JavaScript widget without a stable public data endpoint.
• Michigan — publishes scanned per-notice PDFs only.
• Washington — ESD's "WARN database" page currently ships no data table (JS app removed/broken at source).
• Remaining states are a mix of PDF-only, JS-only, and no-portal situations. Ask in the Issues tab if you need a specific state — portals change, and coverage is re-verified with every update.

If a covered state changes its portal, that state fails in isolation: the run continues, you are only charged for records actually returned, and the run's key-value store contains a COVERAGE report showing exactly which states succeeded and failed.

How to use it

All covered states, notices from the last 90 days:

{
    "noticeDateFrom": "2026-05-14",
    "maxResults": 10000
}

Big layoffs in Texas and Florida this year:

{
    "states": ["TX", "FL"],
    "noticeDateFrom": "2026-01-01",
    "minEmployees": 250,
    "maxResults": 5000
}

Track one company everywhere:

{
    "companySearch": "amazon",
    "noticeDateFrom": "2025-01-01"
}

Leave states empty for all covered states. Without noticeDateFrom, the current + previous calendar year are fetched. maxResults caps the run — you are only charged for records actually returned.

Date semantics (worth 20 seconds)

• noticeDate is the date the state received/posted the notice; layoffDate is when the layoff/closure takes effect (first date, when a range was filed).
• NJ and PA publish only the posting month, not an exact day. For those states noticeDate is null, noticeMonth (e.g. 2026-08) is set, and date filters match by month.
• A handful of filings state effective dates as prose ("9/6/2026 or within 14 days after"); the ISO field is then null and the prose is preserved in layoffDateRaw.

Why not scrape the state sites yourself?

You can — every source this Actor uses is public and official. What you'd be signing up to maintain:

• 10 different formats: xlsx workbooks (3 different layouts), CSV exports with rotating filenames, Google Sheets discovered by scraping a CMS page, paginated ASP.NET tables, and an AEM accordion page.
• Excel serial dates, MM-DD-YY vs MM/DD/YYYY, "thru" ranges, month-only dates, and "or within 14 days after" prose — normalized here into ISO dates plus raw originals.
• Portals that move: three states relocated or restructured their WARN pages in 2026 alone. This Actor re-verifies sources on every update and degrades gracefully per state.

Pricing

Pay per result: $0.005 per notice ($5.00 per 1,000). You are charged per notice actually returned, plus Apify's small platform events: a fraction-of-a-cent Actor start event and $0.00001 per dataset item — about 0.2% on top of the per-notice price. No subscription; a run that finds nothing costs only the fraction-of-a-cent start event.

FAQ

Is this legal? Yes. WARN notices are statutory public records published by state governments precisely so that workers, communities, and businesses can act on them. This Actor reads only official government sources.

How fresh is the data? As fresh as the states publish: CA updates twice a week, most HTML/CSV states update as notices arrive, IL publishes monthly files. Schedule the Actor daily or weekly to keep a rolling feed.

Why is a notice missing that I saw in the news? Either it was filed in a non-covered state (see the table), it is newer than the state's last portal update, or the employer wasn't required to file (WARN thresholds: generally 100+ employees, 50+ affected).

Can I get notified of new notices? Run the Actor on an Apify Schedule with a rolling noticeDateFrom and wire the dataset to email/Slack/webhooks via Apify integrations.