EPA TSCA CDR Chemical Site Screener

Which TSCA chemicals are manufactured or imported at a given site, in what production volume, by whom, and with what worker-exposure and physical-form context — from EPA's 2024 Chemical Data Reporting (CDR) public bulk file.

CDR is a different record class from anything else on the Store. TRI tells you what a facility released. RCRA tells you what it discarded. CDR tells you what it makes and brings into the country — 50,116 site-chemical records covering 4,527 sites and 8,613 substances, with parent-company ownership, production volumes in pounds, worker-exposure bands and physical form.

The one thing that makes this different from a CSV dump

43.15% of 2023 production volumes in this file are the literal string CBI — a Confidential Business Information claim, not a number. Another 13.3% are blank. A build that runs Number(x) || 0 over that column fabricates 28,278 zero-production sites, and does it silently, on a green run.

This actor treats four states as four different facts, and never collapses them:

production_volume_2023_status	Meaning	production_volume_2023_lb
reported	The submitter published a volume	a number (which may legitimately be 0)
cbi_withheld	The submitter claimed it confidential	null
not_known	"Not Known or Reasonably Ascertainable"	null
not_reported	Blank for this record	null

A live drift gate asserts this before a single billable row is produced: one probe pins the national CBI share inside a band, and a second pins a specific record — Shell's Norco, Louisiana refinery, CASRN 7704-34-9 (Sulfur) — where the 2023 production volume is withheld while the 2023 export volume (95,272,804 lb) is published on the same row. If a future change ever turned a withheld volume into a zero or a blank, that probe goes red and the run fails without billing.

Who it is for

• Phase I ESA consultants and environmental due-diligence teams — a CDR filer at an address is a chemical manufacturer or importer at that address. Cross-references by EPA FRS ID into the EPA Contaminated Site Screener and EPA TRI Facility Release Rollup.
• Chemical sourcing and supply-chain analysts — reverse-map a CASRN to every US site that makes or imports it, with domestic and foreign parent companies and D&B numbers.
• EHS, product stewardship and TSCA regulatory affairs — worker-exposure bands, maximum-concentration bands, physical forms and percent byproduct per site-chemical record.
• PFAS and chemical-exposure litigation support — who was making or importing a substance, where, and in what volume band, over 2020–2023.
• CRE lenders and insurance underwriters — chemical manufacturing intensity around a collateral address.

Modes

{
  "mode": "radius",
  "assets": [
    { "lat": 40.8249, "lon": -111.9238, "radiusMiles": 3, "label": "Salt Lake City refinery corridor" },
    { "lat": 29.99435, "lon": -90.40726, "radiusMiles": 3, "label": "Norco LA petrochemical corridor" }
  ],
  "maxResults": 150
}

{ "mode": "chemical", "chemicalIds": ["7704-34-9"], "maxResults": 200 }

{ "mode": "state", "states": ["CO"], "chemicalFilter": [], "maxResults": 250 }

mode is required, so a bare {} call is rejected before the run starts and nothing is billed. maxResults defaults to 200 in code, never as an input-schema default, because Apify injects schema defaults server-side into every run.

What an empty answer means, and what it does not

Every run emits one query_summary row per scope carrying the denominators, and a plain-language coverage_note. This matters because a location screen structurally cannot see 19.3% of this file:

Coordinate state	Rows	Why
Usable coordinate	40,466 (80.74%)	EPA published a real point
Whole site identity CBI	7,074 (14.12%)	Name, address, state, coordinates and FRS ID all withheld
Published as exactly 0.0 / 0.0	2,571 (5.13%)	A null-island sentinel, never a real location
Blank or unparseable	5	—

So an empty radius result means no CDR site with published coordinates reported here — not that no chemical is manufactured or imported at that location. The chemical mode still returns the confidential-identity sites (flagged site_identity_confidential: true), because the substance and its volumes are public even when the site is not.

Related actors on this Store

Join on epa_frs_id:

• EPA Contaminated Site Screener — Superfund, RCRA, ECHO enforcement and tanks at the same coordinates.
• EPA TRI Facility Release Rollup — reported releases to air, water and land, which CDR does not cover.
• EPA RCRA Hazardous Waste Generator Rollup — hazardous waste generated, by federal waste code.
• EPA Drinking Water Quality Screener — SDWA violations and PFAS occurrence for the serving water system.
• Site Due Diligence Bundle — twenty regulatory layers for one coordinate in one call.

Use as an MCP tool

This actor is callable by AI agents through Apify's hosted MCP server. Billing is unchanged when it is called as a tool, and a run that cannot answer fails without billing.

{
  "mcpServers": {
    "apify": {
      "command": "npx",
      "args": ["-y", "@apify/actors-mcp-server", "--actors", "malonestar/epa-tsca-cdr-chemical-site-screener"],
      "env": { "APIFY_TOKEN": "YOUR_APIFY_TOKEN" }
    }
  }
}

Worked agent prompt:

Using the EPA TSCA CDR screener, find every US site that manufactures or imports sulfur (CASRN 7704-34-9). For each site, tell me the parent company, the state, and the 2023 production volume — and be explicit about which volumes are confidential rather than treating them as zero.

Chaining: run this actor first to identify the chemical manufacturers near a coordinate, then pass the epa_frs_id values to EPA Contaminated Site Screener and EPA TRI Facility Release Rollup for the same facilities' regulatory and release history.

Pricing

Pay per result: $8 per 1,000 rows on the standard Free/Bronze/Silver/Gold/Platinum/Diamond tier ladder. One emitted row is one billable result. A run that fails the drift gate emits nothing and bills nothing.

maxResults is a hard cap on the total number of rows a run emits, which is also the hard cap on the bill. It counts every row, including the one query_summary row that each asset, chemical or state produces — those are reserved off the top of the cap and the remainder is spent on site-chemical records. It is a whole-run cap, never per scope, so adding assets never multiplies the bill.

Because each scope costs at least one row, maxResults must be at least the number of scopes in the request. A run asking for more scopes than its cap fails before emitting anything rather than silently answering only some of them — an asset with no row would be indistinguishable from an asset where nothing was found. assets and chemicalIds are capped at 100 entries per run and states at 60, so the scope count is bounded by the input schema as well as by the code.

Every query_summary row publishes the arithmetic: run_max_results_cap, run_scope_summary_rows and run_record_row_budget.

How it fails

Situation	Outcome
mode omitted	HTTP 400 at run creation; the run never starts
A mode's companion field is missing or malformed	Actor.fail with the exact field named and an example value
www.epa.gov is unreachable, returns HTML at HTTP 200, or serves a truncated body	Actor.fail; not a finding of "no chemicals here"
A required column, a zip entry or a canary record changes	Actor.fail naming the probe
The CBI sentinel share collapses	Actor.fail; the actor will not publish a volume it cannot vouch for
A corroborating probe cannot complete	The run proceeds, drift_gate_status reads verified_degraded, and drift_gate_note says which probe was skipped
The query is valid and nothing matched	SUCCESS with a query_summary row explaining the coverage

Set simulateOutage to any value other than none to exercise a failure path deliberately.

Source and vintage

https://www.epa.gov/system/files/other-files/2026-04/2024-cdr-public-csv-data.zip (14,378,975 bytes, Last-Modified: Mon, 20 Apr 2026 22:03:49 GMT), keyless, linked from EPA's Access CDR Data page.

EPA replaces these files in place under an unchanged URL, so the vintage is read from Last-Modified on every run and published on every row as source_last_modified and source_vintage_status. Never trust a hard-coded CDR vintage.

FAQ

What is TSCA Chemical Data Reporting? Every four years, manufacturers and importers of TSCA-listed chemicals above a volume threshold must report production volumes, site information and use context to EPA. The 2024 cycle covers calendar years 2020–2023.

Does this include chemical releases or waste? No. CDR is manufacture and import. Use the EPA TRI Facility Release Rollup for releases and the EPA RCRA Hazardous Waste Generator Rollup for waste.

Why is a production volume null instead of 0? Because the submitter claimed it confidential, reported it as not known, or left it blank. Read production_volume_2023_status — a real reported zero is a number, and it is a different fact.

Why can't I find a site I know files under CDR? 7,074 records withhold the entire site identity as CBI, and 2,571 more are published at coordinates 0.0 / 0.0. Query by chemical mode to see the confidential-identity records; they carry the substance and volumes without a location.

Are the production volumes in pounds? Yes. EPA's own data dictionary states all CDR quantities are reported in pounds; production_volume_units says so on every row.

Why does workers_exposed_published_label look wrong? Because it is. EPA publishes two different label texts for the same band code and writes a literal ? where a >= sign belongs. This actor decodes from the code, and republishes EPA's own text verbatim beside it so nothing is hidden.

Output fields

Every row carries record_type. site_chemical rows are the reported records; one query_summary row per scope carries the coverage the answer depends on.

Query scope

Field	Type	Description
record_type	string or null	site_chemical for a reported site-chemical record, query_summary for the one coverage row emitted per query scope.
query_mode	string or null	The mode this row was produced under: radius, chemical or state.
query_scope_label	string or null	Label of the scope: the asset label, the CASRN, or the state code.
asset_label	string or null	Caller-supplied label for the screened coordinate. Null outside radius mode.
asset_latitude	number or null	Latitude of the screened coordinate. Null outside radius mode.
asset_longitude	number or null	Longitude of the screened coordinate. Null outside radius mode.
asset_radius_miles	number or null	Search radius in miles. Null outside radius mode.
distance_miles	number or null	Great-circle distance from the screened coordinate to the site, in miles. Null when the match was not made on coordinates.
match_basis	string or null	How this record was matched: coordinate_radius, chemical_id, site_state, or no_coordinate_included_on_request.

Site identity

Field	Type	Description
site_name	string or null	Published site name with any trailing asterisk marker removed. Null when EPA withholds the site identity as confidential.
site_name_raw	string or null	Site name exactly as EPA published it, including the trailing asterisk that appears on 108 rows and is not explained in EPA's own data dictionary.
site_name_has_asterisk_marker	boolean or null	True when the published site name carries the undocumented trailing asterisk; null when the name is not disclosed.
site_name_status	string or null	reported, cbi_withheld, not_known or not_reported for the site name.
site_address_line1	string or null	Street address of the reporting site. Null when withheld or absent.
site_city	string or null	City of the reporting site.
site_county	string or null	County or parish of the reporting site.
site_state	string or null	Two-letter state code of the reporting site. Null when the site identity is confidential.
site_state_status	string or null	reported, cbi_withheld, not_known or not_reported for the site state.
site_postal_code	string or null	ZIP code of the reporting site.
site_latitude	number or null	Latitude EPA published for the site. Null when confidential, absent, or published as the 0/0 sentinel.
site_longitude	number or null	Longitude EPA published for the site, under the same rules as site_latitude.
site_coordinate_status	string or null	disclosed, cbi_withheld, not_reported, or null_island_sentinel when EPA published exactly 0/0 (2,571 rows nationally).
site_identity_confidential	boolean or null	True when EPA withheld the whole site identity as CBI, which is why the row carries a chemical but no location.
epa_tsca_program_id	string or null	EPA TSCA program identifier for the site.
epa_frs_id	string or null	EPA Facility Registry Service identifier. This is the join key into EPA Contaminated Site Screener and EPA TRI Facility Release Rollup.
site_dun_bradstreet_number	string or null	Dun & Bradstreet number of the site.
site_naics_code_1	string or null	Primary NAICS code of the site, split out of the packed code-plus-description string EPA publishes.
site_naics_description_1	string or null	Primary NAICS description.
site_naics_activity_1	string or null	The CDR activity EPA associates with the primary NAICS, e.g. Manufacture or Import.
site_naics_code_2	string or null	Second NAICS code, populated on about a quarter of records.
site_naics_description_2	string or null	Second NAICS description.
site_naics_code_3	string or null	Third NAICS code, populated on about a fifth of records.
site_naics_description_3	string or null	Third NAICS description.

Corporate parentage

Field	Type	Description
standardized_parent_company	string or null	EPA-standardised parent company name. Null when the corporate block is confidential.
parent_company_status	string or null	reported, cbi_withheld, not_known or not_reported for the parent company block.
domestic_parent_company	string or null	Domestic parent company name.
domestic_parent_city	string or null	Domestic parent company city.
domestic_parent_state	string or null	Domestic parent company state.
domestic_parent_postal_code	string or null	Domestic parent company ZIP code.
domestic_parent_dun_bradstreet_number	string or null	Domestic parent company Dun & Bradstreet number.
foreign_parent_company	string or null	Foreign parent company name where one is reported.
foreign_parent_city	string or null	Foreign parent company city.
foreign_parent_country_code	string or null	Foreign parent company country code.
foreign_parent_dun_bradstreet_number	string or null	Foreign parent company Dun & Bradstreet number.
has_foreign_parent	boolean or null	True when EPA published a foreign parent, false when it published the corporate block without one, and null when the corporate block is confidential so foreign ownership is unknown rather than absent.

Chemical identity

Field	Type	Description
chemical_name	string or null	Systematic chemical substance name as listed on the public portion of the TSCA Inventory.
chemical_id	string or null	CASRN, or an EPA Accession Number when the specific chemical identity is confidential.
chemical_id_no_dashes	string or null	The same identifier with dashes stripped, as EPA publishes it.
chemical_id_type	string or null	CASRN or Accession Number.
chemical_identity_confidential	boolean or null	True when the identifier is an Accession Number, which per EPA means the chemical identity sits on the confidential portion of the TSCA Inventory.

Reported activity

Field	Type	Description
activity	string or null	Manufacture, Import or Both. Null when the activity itself is claimed confidential.
activity_status	string or null	reported, cbi_withheld, not_known or not_reported for the activity.
is_manufacturer	boolean or null	True when the site manufactures the substance, false when it only imports it, null when the activity is withheld.
is_importer	boolean or null	True when the site imports the substance, false when it only manufactures it, null when the activity is withheld.
chemical_never_physically_at_site	boolean or null	True when the submitter reported the substance is never physically at the site (imported and shipped onward). Null when withheld, not known, or not reported.
chemical_never_physically_at_site_status	string or null	reported, cbi_withheld, not_known or not_reported for the never-at-site question.

Production volumes (CBI-critical)

Field	Type	Description
production_volume_2023_lb	number or null	Total 2023 production volume in pounds. NULL, never zero, when the value is confidential, not known, or not reported.
production_volume_2023_status	string or null	reported, cbi_withheld, not_known or not_reported. This is the field that separates a withheld volume from a real zero.
production_volume_2023_cbi_withheld	boolean or null	True when the 2023 production volume was claimed confidential.
domestic_production_volume_2023_lb	number or null	Domestically manufactured portion of the 2023 volume, in pounds.
domestic_production_volume_2023_status	string or null	Disclosure status of the domestic 2023 volume.
import_volume_2023_lb	number or null	Imported portion of the 2023 volume, in pounds.
import_volume_2023_status	string or null	Disclosure status of the 2023 import volume.
production_volume_2022_lb	number or null	2022 production volume in pounds.
production_volume_2022_status	string or null	Disclosure status of the 2022 volume.
production_volume_2021_lb	number or null	2021 production volume in pounds.
production_volume_2021_status	string or null	Disclosure status of the 2021 volume.
production_volume_2020_lb	number or null	2020 production volume in pounds.
production_volume_2020_status	string or null	Disclosure status of the 2020 volume.
volume_used_on_site_2023_lb	number or null	Volume used on site in 2023, in pounds.
volume_used_on_site_2023_status	string or null	Disclosure status of the volume used on site.
volume_exported_2023_lb	number or null	Volume exported in 2023, in pounds.
volume_exported_2023_status	string or null	Disclosure status of the exported volume.
production_volume_units	string or null	Unit for every volume field. CDR reports all quantities in pounds.
production_volume_disclosure	string or null	fully_reported, partially_reported, partially_cbi_withheld, fully_cbi_withheld or not_reported across the four annual volumes.
production_volume_change_2020_to_2023_lb	number or null	Change in pounds from 2020 to 2023, computed only when both endpoints are genuinely reported.
production_volume_change_2020_to_2023_pct	number or null	Percentage change from 2020 to 2023, computed only when both endpoints are genuinely reported.
production_volume_trend_status	string or null	computed, or not_computable when an endpoint is withheld or missing.
production_volume_trend_note	string or null	Which endpoint prevented the trend from being computed.

National aggregate context

Field	Type	Description
national_aggregated_pv_2023_text	string or null	EPA's nationally aggregated 2023 production volume for the substance, exactly as published. It is a BAND on 46,141 of 50,116 rows, not a number.
national_aggregated_pv_2023_low_lb	number or null	Lower bound of the national band in pounds; equal to the upper bound when EPA published a single number.
national_aggregated_pv_2023_high_lb	number or null	Upper bound of the national band in pounds.
national_aggregated_pv_2023_is_range	boolean or null	True when EPA published a band, false when it published a single number, null when it published nothing.
national_aggregated_pv_2023_status	string or null	Disclosure status of the national 2023 aggregate.
national_aggregated_pv_2022_text	string or null	National aggregate for 2022 as published.
national_aggregated_pv_2022_status	string or null	Disclosure status of the national 2022 aggregate.
national_aggregated_pv_2021_text	string or null	National aggregate for 2021 as published.
national_aggregated_pv_2021_status	string or null	Disclosure status of the national 2021 aggregate.
national_aggregated_pv_2020_text	string or null	National aggregate for 2020 as published.
national_aggregated_pv_2020_status	string or null	Disclosure status of the national 2020 aggregate.

Exposure and physical form

Field	Type	Description
workers_exposed_code	string or null	EPA band code W1-W8 for workers reasonably likely to be exposed, or CBI/NKRA as published.
workers_exposed_band	string or null	Plain-language band decoded from the code, not from EPA's published label text.
workers_exposed_low	number or null	Lower bound of the worker band.
workers_exposed_high	number or null	Upper bound of the worker band; null for the open-ended top band.
workers_exposed_published_label	string or null	EPA's own label text, republished verbatim. Note that EPA publishes two different labels for the same code and writes a literal question mark where a greater-than-or-equal sign belongs.
workers_exposed_status	string or null	reported, cbi_withheld, not_known or not_reported for worker exposure.
max_concentration_code	string or null	EPA band code M1-M5 for maximum concentration, or CBI/NKRA as published.
max_concentration_band	string or null	Plain-language concentration band decoded from the code.
max_concentration_low_pct	number or null	Lower bound of the concentration band in percent.
max_concentration_high_pct	number or null	Upper bound of the concentration band in percent.
max_concentration_published_label	string or null	EPA's own concentration label, republished verbatim.
max_concentration_status	string or null	reported, cbi_withheld, not_known or not_reported for maximum concentration.
percent_byproduct_code	string or null	EPA band code B1-B4 for percent byproduct. 481 rows carry a raw percentage here instead of a band code.
percent_byproduct_band	string or null	Plain-language byproduct band decoded from the code.
percent_byproduct_low_pct	number or null	Lower bound of the byproduct band in percent.
percent_byproduct_high_pct	number or null	Upper bound of the byproduct band in percent.
percent_byproduct_published_label	string or null	EPA's own byproduct label, republished verbatim.
percent_byproduct_status	string or null	reported, cbi_withheld, not_known or not_reported for percent byproduct.
percent_byproduct_code_is_numeric	boolean or null	True when EPA published a raw percentage in the band-code column rather than a B-code.
physical_forms	array or null	Physical forms of the substance at the site, split from EPA's semicolon-delimited list. Null when withheld or not reported.
physical_forms_status	string or null	reported, cbi_withheld, not_known or not_reported for physical form.
physical_form_pct_dry_powder	number or null	Share of the production volume in dry powder form, as a percentage. Despite EPA naming these columns "PF PPV" they hold percentages, not volumes.
physical_form_pct_pellets_or_large_crystals	number or null	Share of the production volume as pellets or large crystals, in percent.
physical_form_pct_wet_solid	number or null	Share of the production volume as water- or solvent-wet solid, in percent.
physical_form_pct_other_solid	number or null	Share of the production volume as other solid, in percent.
physical_form_pct_gas_or_vapor	number or null	Share of the production volume as gas or vapor, in percent.
physical_form_pct_liquid	number or null	Share of the production volume as liquid, in percent.
physical_form_pct_nkra	number or null	Share of the production volume whose physical form is not known or reasonably ascertainable, in percent.
physical_form_pct_status	string or null	Disclosure status of the physical-form percentage block.
recycled_or_otherwise_used	boolean or null	True when the substance is recycled or otherwise used at the site. EPA publishes this in mixed case, which is normalised here.
recycled_or_otherwise_used_status	string or null	reported, cbi_withheld, not_known or not_reported for the recycling question.

Coverage and denominators

Field	Type	Description
cdr_reporting_cycle	number or null	CDR submission cycle this file covers.
cdr_principal_reporting_year	number or null	The reporting year for which site-level volumes are published in full detail.
national_rows_total	number or null	Total site-chemical records in the national Manufacture-Import table, excluding blank rows.
national_distinct_sites	number or null	Distinct sites with a disclosed TSCA program identifier nationally.
national_distinct_chemicals	number or null	Distinct chemical identifiers nationally.
national_site_identity_cbi_rows	number or null	National count of records whose entire site identity is confidential, and which therefore cannot appear in any location or state query.
national_coordinate_usable_rows	number or null	National count of records with a usable coordinate.
national_coordinate_null_island_rows	number or null	National count of records published at exactly 0/0.
national_coordinate_missing_rows	number or null	National count of records with a blank or unparseable coordinate.
national_coordinate_usable_share_pct	number or null	Share of national records carrying a usable coordinate, after excluding confidential identities and the 0/0 sentinel.
national_pv_2023_cbi_rows	number or null	National count of confidential 2023 production volumes.
national_pv_2023_reported_rows	number or null	National count of disclosed 2023 production volumes.
national_pv_2023_not_reported_rows	number or null	National count of blank 2023 production volumes.
national_pv_2023_not_known_rows	number or null	National count of 2023 production volumes reported as not known or reasonably ascertainable.
national_pv_2023_cbi_share_pct	number or null	Share of national records whose 2023 production volume is confidential.
scope_matched_rows_total	number or null	How many records matched this scope before the maxResults cap.
scope_rows_emitted	number or null	How many records were emitted for this scope after the cap.
scope_results_truncated	boolean or null	True when the maxResults cap removed matching records.
scope_rows_invisible_to_location_screen	number or null	Records excluded from this radius screen because EPA published no usable coordinate for them.
scope_rows_with_site_identity_cbi	number or null	Matched records whose site identity is confidential, so they carry a chemical but no location.
scope_distinct_sites	number or null	Distinct sites emitted for this scope.
scope_distinct_chemicals	number or null	Distinct chemicals emitted for this scope.
run_max_results_cap	number or null	The maxResults value this run was given. It is the hard cap on total emitted rows, and therefore on the bill, counting query_summary rows as well as site-chemical rows.
run_scope_summary_rows	number or null	How many rows of the maxResults cap were reserved for query_summary rows - exactly one per asset, chemical or state in the request. This row is one of them.
run_record_row_budget	number or null	The rows left for site-chemical records after the query_summary rows were reserved: maxResults minus run_scope_summary_rows.
source_blank_rows_skipped	number or null	Completely blank rows present in EPA's CSV and skipped, counted rather than silently dropped.
coverage_note	string or null	Plain-language statement of what this query could and could not see, so an empty result is not read as an absence of chemical activity.

Provenance and per-source status

Field	Type	Description
source_file_url	string or null	The EPA bulk file this run read.
source_last_modified	string or null	Last-Modified header of the bulk file, read on every run because EPA replaces these files in place under an unchanged URL.
source_content_length_bytes	number or null	Size in bytes of the bulk file this run read.
source_vintage_status	string or null	current, stale_over_4_years, future_dated, unparseable or unknown, derived from Last-Modified.
retrieved_at	string or null	UTC timestamp when this run fetched the file.
cdr_bulk_file_status	string or null	Per-source status for the EPA bulk file.
cdr_manufacture_import_status	string or null	Per-source status for the Manufacture-Import table.
drift_gate_status	string or null	verified when every probe passed, or verified_degraded when a corroborating probe could not complete and the run proceeded.
drift_probes_verified	number or null	How many live drift probes passed before any row was billed.
drift_probes_unavailable	number or null	How many drift probes could not complete.
drift_gate_note	string or null	Which corroborating probes could not complete, when the gate ran degraded.
source_row_index	number or null	Line number of this record in EPA's CSV, for reproducibility.