What we do

WARN Firehose aggregates US labor market data from six federal sources into a single, queryable platform:

Records are scraped daily, normalized, deduplicated, and surfaced via a REST API, an MCP server for AI agents, CSV/JSON/Parquet exports, alerts, and 12,000+ SEO pages for organic discovery.

Why we built this

WARN Act notices are public records, but they are scattered across 50 separate state workforce agency websites, each with different PDF/HTML/Excel formats, naming conventions, and update cadences. Until recently, nobody had quietly and reliably consolidated them. The same is true for the four other federal datasets we cover — each lives in its own portal, schema, and refresh cadence, making cross-dataset analysis prohibitively expensive for most analysts.

We saw three audiences underserved by the status quo:

WARN Firehose is our answer. It is independent, audit-friendly, and built to outlast trend cycles.

Who we are

Kamal Sharma
Founder & engineer. Built warnfirehose.com from a one-state Python scraper into a six-dataset platform with 12K+ public pages and a paid API tier. Background in software engineering and data infrastructure. Reachable at [email protected] or via the contact form.

WARN Firehose is independently owned and operated. No external investors, no editorial influence from labor agencies, no syndication deals that affect what we publish or how. Every decision — what data to ingest, how to normalize it, what to charge — is documented in the public GitHub repository.

How we work

By the numbers

85K+
WARN notices
5.9M
H-1B/LCA petitions
42 yrs
unemployment data
50
states covered (47 WARN)
Daily
refresh cadence
12K+
indexed public pages

Contact & press

Press inquiries, data requests, or partnership questions: [email protected]. Replies usually within one business day. Journalists get free API access — details on the Press Room.

Looking for our data sources, scraping cadence, and known gaps? Read our methodology. Looking for press citations and academic references? See Cited By.