API.md

A free WARN Act API — 61,431 layoff notices, 48 states, one GET request

No key, no signup, no rate limit, CORS open. Every US state publishes the mass-layoff notices employers file under the WARN Act, and every state publishes them differently. We scrape all 48 of them every morning, normalize them into one 11-field schema, resolve the employer names, and serve the result as plain JSON and CSV over HTTPS. Rebuilt 2026-09-25; newest notice in this build is dated 2026-09-25. CC BY 4.0 — credit "WARN Feed".

Full docs, examples and a per-state list: https://approjects-warn-act-notices.static.hf.space/api.html

Quickstart

# the whole archive as JSON
curl -sL https://approjects-warn-act-notices.static.hf.space/data/warn_notices.json | jq '.[0]'
# just what is new — poll this one daily
curl -sL https://approjects-warn-act-notices.static.hf.space/data/latest.json | jq '.count, .generated_at'
# poll politely: every file carries an ETag; send it back and an unchanged file is a 304 with no body
curl -sL -o latest.json -D - -H 'If-None-Match: "<etag from last run>"' https://approjects-warn-act-notices.static.hf.space/data/latest.json
# streaming: one notice per line (NDJSON), no 20 MB buffer
curl -sL https://approjects-warn-act-notices.static.hf.space/data/warn_notices.ndjson | head -n 3 | jq -c '{company_canonical, state, notice_date}'
# one state, as CSV
curl -sL https://approjects-warn-act-notices.static.hf.space/data/by-state/ca.csv | head -3

Files over ~10 MB (the JSON and NDJSON archives) answer 302 to a CDN copy that sends the same CORS header — keep -L (every HTTP library follows by default).

import pandas as pd
df = pd.read_csv("https://approjects-warn-act-notices.static.hf.space/data/latest.csv")
print(len(df), df.columns.tolist())

From an LLM or an agent: the MCP server

warn-mcp wraps these endpoints in six read-only Model Context Protocol tools, so an agent asks a question instead of pulling a 9 MB file into a context window: employer + state + date-range search, the rolling new-notice feed, one employer's whole history back to 1988, monthly state totals, the revision log (which rows changed or vanished between daily builds, tagged by what is mechanically distinguishable — never a guess at why), and a freshness call. Standard library only, no key, disk cache with ETag revalidation.

claude mcp add warn -- uvx --from git+https://github.com/APVentureEngine/warn-mcp warn-mcp

Any client that reads an mcpServers block (Claude Desktop, Cursor, Continue):

{"mcpServers": {"warn": {"command": "uvx",
  "args": ["--from", "git+https://github.com/APVentureEngine/warn-mcp", "warn-mcp"]}}}

Prefer a URL to a subprocess — ChatGPT and Claude.ai connectors, n8n, anything that only speaks remote MCP? The same six tools are hosted over Streamable HTTP at https://mQpACuIiQcnwK93JY.apify.actor/mcp. One honest caveat: the host meters compute to whoever calls it, so that endpoint needs YOUR own free Apify API token in an Authorization: Bearer header. We charge nothing for it and keep nothing. The install-free bundle above needs no token of any kind, so use it instead if your client can.

Both routes are listed as io.github.APProj/warn-mcp in the official MCP registry.

Source, tool reference and issues: https://github.com/APVentureEngine/warn-mcp

Endpoints

Endpoint	What	Records	Size
/data/warn_notices.json	Whole archive, JSON	61,431 notices	20.1 MB
/data/warn_notices.ndjson	Whole archive, NDJSON	61,431 notices	18.7 MB
/data/latest.csv	Recent notices, CSV	88 notices	16 kB
/data/latest.json	Recent notices, JSON	88 notices	36 kB
/datapackage.json	Schema (Frictionless Table Schema)	11 fields	3 kB
/data/coverage.json	Per-state coverage + freshness	48 states	11 kB
/data/stale.json	Sources behind their usual cadence	10 late sources	2 kB
/data/trends.json	Monthly series	24 months	56 kB
/data/search-index.json	Employer search index	29,971 employers	2.7 MB
/data/badge.json	Shields.io endpoint badge	—	153 B
/feed.xml	RSS of new notices	50 items	29 kB

Per-state CSVs for 48 states: https://approjects-warn-act-notices.static.hf.space/data/by-state/<two-letter-code-lowercase>.csv (AK, AL, AZ, CA, CO, CT, DC, DE, FL, GA, HI, IA, ID, IL, IN, KS, KY, LA, MA, MD, ME, MI, MN, MO, MS, MT, NC, ND, NE, NJ, NM, NV, NY, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VA, VT, WA, WI, WV).

Schema

Field	Type	Meaning
id	string	Stable dedupe id (source record number where available, else content hash)
state	string	2-letter US postal code
company	string	Employer name exactly as published by the state
company_canonical	string	Cleaned employer name; groups the same employer across states, renotices and store numbers
company_dba	string	Trade name when the state published a dba/aka alias
location	string	City/county/address, best effort
employees_affected	integer	Workers affected; empty if the state omitted it
notice_date	date	ISO date the notice was received/posted (NJ: month precision, day pinned to 01)
effective_date	date	ISO date the layoff/closure takes effect
notice_type	string	Layoff/closure label as published
first_seen	datetime	UTC timestamp the pipeline first saw this notice

What this API does not do

• It does not push. A snapshot cannot tell you that your employer just filed; you would poll and diff it yourself. warn-layoff-alert-bot is a 120-line, dependency-free example that does exactly that from a daily GitHub Actions cron and posts to Slack or Discord (MIT; fork it).
• It does not filter. Static files have no query parameters.
• It is a once-a-morning rebuild, not a stream.

If what you wanted was an alert rather than a file, WARN Watch runs your employer list (up to 500 names, or whole states) against every daily refresh for a year and pushes hits to a private page, RSS, a calendar and your Slack/Discord/Teams webhook — $49/year, one payment, or a free 30-day trial.

Need an endpoint we do not publish? Open an issue and say which field, state or shape — the pipeline runs every day anyway.