US Industrial Facility Data — EPA ECHO Compliance & Enforcement

EPA ECHO is the compliance record of every regulated industrial site in the United States — factories, refineries, treatment plants, waste handlers. Query it by company name and you get their facilities, what industry they are in, whether the EPA has found violations, inspected them or fined them, their greenhouse-gas and toxic-release figures — for the facilities that report to the Toxics Release Inventory, the parent company and the public contact they filed — and, from OSHA's injury-reporting data, how many people work at the site and the employer's EIN.

No API key · Official source · Pay only for delivered rows · Same schema across the series

What data do you get?

Field	Description
companyName	Facility name as registered with the EPA
registrationNumber	EPA Registry ID (FRS identifier)
status	active when ECHO flags the facility active in at least one EPA program, inactive otherwise
address	Facility street address
city, county, state, postalCode	Facility location
industry	NAICS title of industryCode, from ECHO's own NAICS lookup
industryCode	First NAICS code on the record, falling back to SIC
naicsCodes, sicCodes	All NAICS and SIC codes on the record
employees	Annual average employees at the facility's address, from the operator's OSHA Form 300A (several establishments of the same employer at one site are summed)
employeesAsOf, employeesSource	Calendar year of that figure, and how it was put together
employerName, taxNumber	The employer that files OSHA Form 300A for the site, and its EIN — often a subsidiary (TYSON FARMS, INC. at a Tyson Foods plant)
taxNumberSource	Where taxNumber comes from, and whether it equals the parent company's EIN on SEC EDGAR
parentCompany	Parent company declared on the facility's TRI filing (EPA-standardized name)
parentDuns	D-U-N-S number of that parent, as filed
parentEin, parentTicker, parentCik	EIN, stock ticker and EDGAR CIK of that parent, when it is listed on a US exchange
foreignParentCompany	Foreign parent company, when one is declared
contactName, phone, email	The public contact the facility named on its TRI filing
triFacilityId	TRI facility ID the parent and contact fields come from
complianceStatus	Overall ECHO compliance status
airComplianceStatus	Clean Air Act compliance status
waterComplianceStatus	Clean Water Act compliance status
wasteComplianceStatus	RCRA hazardous waste compliance status
significantNoncompliance	True when flagged a High Priority / Significant / Serious violator
quartersInNoncompliance	Quarters in non-compliance out of the last twelve
inspectionCount	Compliance monitoring activities in the last five years
lastInspectionOn	Date of the most recent inspection
formalActionCount	Formal enforcement actions in the last five years, all programs
penaltyCount	Penalties assessed in the last five years
penaltyTotal	Total penalties assessed in the last five years, all programs, USD
lastPenaltyOn, lastPenaltyAmount	The most recent penalty on record and its amount, USD
ghgEmissionsTonsCo2e	Greenhouse-gas emissions reported to the EPA GHG program, metric tons CO2e per year
triReleasesLbs	Toxic chemical releases reported to TRI for the latest reporting year, pounds
latitude, longitude	Facility coordinates
sourceUrl	ECHO detailed facility report

Every row also carries query (what you asked for), scrapedAt (ISO timestamp) and, when a lookup fails, error explaining why.

What the EPA does not publish. A row here is a facility, not a company record, so legalForm, incorporatedOn, dissolvedOn, capital, website and officers — part of the company schema shared across this series — stay null. employees and taxNumber come from OSHA (see Notes and limits).

Example output

A real row for Tesla's Fremont plant (contact details masked here; the dataset carries them in full):

{
  "country": "US",
  "registry": "EPA ECHO",
  "companyName": "TESLA, INC.",
  "registrationNumber": "110000482898",
  "status": "active",
  "address": "45500 FREMONT BOULEVARD, FREMONT, CA, 94538",
  "city": "FREMONT",
  "county": "ALAMEDA",
  "state": "CA",
  "postalCode": "94538",
  "industry": "Automobile and Light Duty Motor Vehicle Manufacturing",
  "industryCode": "336110",
  "naicsCodes": ["336110", "336111", "33611", "336399"],
  "sicCodes": ["3711"],
  "taxNumber": "91-2197729",
  "employees": 16609,
  "employerName": "Tesla",
  "taxNumberSource": "OSHA Form 300A; the same EIN as the parent company on SEC EDGAR",
  "employeesAsOf": "2025",
  "employeesSource": "OSHA Form 300A annual average, calendar year 2025",
  "parentCompany": "TESLA INC",
  "parentEin": "91-2197729",
  "parentTicker": "TSLA",
  "contactName": "V••••• O•••••",
  "phone": "925-•••-••••",
  "email": "v•••••@tesla.com",
  "triFacilityId": "94538NWNTD45500",
  "complianceStatus": "Significant Violation",
  "airComplianceStatus": "Violation Unaddressed; Local Has Lead Enforcement",
  "waterComplianceStatus": "No Violation Identified",
  "wasteComplianceStatus": "No Violation Identified",
  "significantNoncompliance": true,
  "quartersInNoncompliance": 12,
  "inspectionCount": 8,
  "lastInspectionOn": "2025-07-23",
  "formalActionCount": 1,
  "penaltyCount": 1,
  "penaltyTotal": 275000,
  "lastPenaltyOn": "2022-02-17",
  "lastPenaltyAmount": 275000,
  "ghgEmissionsTonsCo2e": 47053,
  "triReleasesLbs": 2125966,
  "latitude": 37.49634,
  "longitude": -121.94826,
  "sourceUrl": "https://echo.epa.gov/detailed-facility-report?fid=110000482898"
}

Input

{
  "queries": ["Tesla","Dow Chemical","Nucor Steel"],
  "maxResultsPerQuery": 10,
  "maxConcurrency": 3,
  "includeRaw": false
}

Input	What it does
queries	Company or facility names (Tesla, Dow Chemical). ECHO matches on the facility name as registered, which is usually the operating entity rather than the corporate parent.
maxResultsPerQuery	Caps how many rows one query may produce.
maxConcurrency	How many queries run at once. Lower it if the source starts throttling.
includeRaw	Attaches the source's untouched record under raw, for fields this actor does not map.
requestDelayMs	Politeness delay between requests.
proxyConfiguration	Optional. ECHO is an open federal API and rarely needs a proxy.
state	Two-letter state code (TX, CA) to search one state only.

What people use it for

• Industrial site prospecting — find every plant a manufacturer runs, with its industry, coordinates and the environmental contact it filed, for territory planning or selling compliance services.
• ESG and supply-chain screening — check a supplier's violation history, penalties, greenhouse-gas emissions and toxic releases before contracting.
• Ownership mapping — the TRI parent company shows who owns a plant today, even when the site still carries a former owner's name (a "Dow Chemical" plant now owned by Occidental, for example).
• Environmental research — assemble a facility panel by industry code and track compliance over time.

Notes and limits

• ECHO is a two-step API — a search returns a query ID, then results are paged from it. The actor does both, so a query costs two requests, plus one run-wide request for the NAICS titles.
• Parent company and contact come from the Toxics Release Inventory, so only facilities that report to TRI carry them — large manufacturers mostly do, offices, service centres and small sites do not. Each such facility costs one more request (EPA Envirofacts); if it fails, the row keeps its ECHO fields and the log says which facility was affected.
• Facilities are registered under operating names, not corporate parents. Searching Tesla finds sites named Tesla; parentCompany then tells you who owns the TRI reporters among them.
• Listed parents come from SEC EDGAR, matched on the exact company name (the SEC's list of exchange-listed companies is read once per run, then one EDGAR record per parent). The EIN is the parent's, not the facility operator's — a plant's own operating company (e.g. NUCOR STEEL DECATUR LLC) has its own.
• Headcount and EIN come from OSHA's Form 300A summaries (Injury Tracking Application), which establishments with 20 or more employees in designated industries, and 250 or more in all others, file every year. A facility is matched at its address, then only when exactly one employer there passes every test: its name starts with the facility's or the parent's brand, its industry sector is the facility's, it is not a contractor on site (cleaning, security, catering, staffing…), and its EIN mostly files under that brand. In a 200-facility test that played no part in tuning the matching, 44.5% of facilities got a headcount (51% when searching company names, 26% for generic words such as Plating or Foundry) and a manual check of every match found no wrong employer. Small sites, offices and non-reporting industries stay empty.
• The EIN is the employer's own as it filed with OSHA (three quarters of the matches carry one); it equals the parent's SEC EIN when the parent runs the plant itself. An EIN one or two digits away from the parent's SEC EIN is a typo in the filing and is left out. The OSHA data is read through a monthly-rebuilt foXLabs index; if it cannot be reached, these fields stay empty and the run goes on.
• ECHO's name search matches loosely and its order is not relevance. The actor ranks up to 1,000 matches per query: names containing your search as whole words first, and among those the active facilities that report to TRI or the greenhouse-gas program — the plants — ahead of offices and service sites.
• Five-year figures (inspectionCount, formalActionCount, penaltyCount, penaltyTotal) follow ECHO's own window. lastFormalActionOn and lastPenaltyOn report the most recent event on record, which can be older.

Where the data comes from

The EPA publishes ECHO as a documented open REST service with no key and no quota; the TRI facility records come from the EPA's Envirofacts data service, also keyless. Sources: EPA Enforcement and Compliance History Online (ECHO), EPA Envirofacts, SEC EDGAR, OSHA Injury Tracking Application data

FAQ

Is this EPA ECHO scraper free?

The data source is free and needs no API key — you pay only for the rows the run delivers ($0.004 each). Failed or empty lookups are never charged.

Do I need an API key or a login?

No. The EPA publishes ECHO and Envirofacts as open REST services with no key.

What can I search by?

By facility or parent company name (Tesla, Dow Chemical). Narrow with the state input, which maps onto ECHO's own state filter.

How current is the data?

Every run queries the source live, so results are as fresh as the source itself. ECHO refreshes from the underlying enforcement systems weekly; contact and parent details are read from the facility's TRI record in Envirofacts at run time.

How fast is it, and how many queries can I run?

Queries run concurrently (3 at a time by default, tunable in the input). A prefilled run finishes in well under a minute; the TRI lookups are what take the time. Large lists scale roughly linearly and stay well inside a normal run timeout.

Can I export the results to CSV, Excel or JSON?

Yes. Apify datasets export to CSV, Excel, JSON, XML and HTML, and can be pulled through the API or pushed to your own storage.

What happens when a query returns nothing?

You still get a row, carrying your original query and an error field explaining why. Nothing is silently dropped, and you are not charged for it.

Is scraping this data legal?

Yes. ECHO is published by the EPA precisely so the public can see the compliance history of regulated facilities, and the TRI public contact is the person a facility designates for public inquiries about its filing.

Changelog

0.1.10 — 2026-09-25 — site headcount and employer EIN (OSHA)

• New: employees, employeesAsOf, employeesSource, employerName, taxNumber, taxNumberSource, from OSHA's Form 300A summaries matched at the facility's address. 44.5% of facilities filled in a 200-facility test that played no part in tuning, no wrong employer found in a manual check of every match.
• Fixed: parentEin showed 00-0000000 for Owens Corning plants. EDGAR carries a placeholder of zeros for some companies; it is no longer passed on as an EIN.
• No pricing change.

0.1.9 — 2026-09-25 — listed parent company: EIN and ticker

• New parentEin, parentTicker, parentCik. When the parent company a facility declares to the Toxics Release Inventory is listed on a US exchange, its EDGAR record gives the EIN and ticker: NUCOR CORP → NUE, 13-1860817; TESLA INC → TSLA. The match is on the exact company name — punctuation and the spelling of the legal suffix are evened out, the suffix itself must agree, and a name two listed companies share matches neither. Subsidiaries and private parents stay empty. 67% of rows in the prefilled run.
• SEC requests identify the actor and a contact address, as the SEC's fair-access policy asks.
• No pricing change.

0.1.8 — 2026-09-25 — dates, coordinates and penalties corrected; industry, parent company and contacts added

• Dates were wrong and are now correct. ECHO writes dates as MM/DD/YYYY and they were read day-first: 10/16/2025 became the impossible 2025-16-10, and 08/02/2002 became February 8 instead of August 2. This affected lastInspectionOn, lastFormalActionOn and lastPenaltyOn in every earlier version (7 of 33 rows carried an impossible date in a reference run of 0.1.7; the others with a day of 12 or less were silently swapped).
• longitude is filled (it was always empty: ECHO does not send it unless asked), and coordinates with exactly three decimals are no longer corrupted: the number parser read 41.389 as 41,389, which put 4 of 33 latitudes out of range in a reference run of 0.1.7.
• penaltyTotal and formalActionCount now cover all programs, as ECHO totals them over five years. They used to count Clean Air Act penalties only, and Clean Air Act (else drinking water) actions only — values rise for facilities fined under the water or waste programs.
• New: industry (NAICS title), parentCompany, parentDuns, foreignParentCompany, contactName, phone, email and triFacilityId (TRI reporters), lastPenaltyAmount, ghgEmissionsTonsCo2e, triReleasesLbs.
• Better matches: up to 1,000 search hits are ranked instead of the first 50, so Tesla now returns the Fremont, Nevada, Texas and Buffalo plants instead of service centres and streets named Tesla Road.
• README price corrected to the actual $0.004 per row (it said $0.005). No pricing change.

---

Built by Fox Labs — B2B company intelligence from public sources, as clean JSON.