US Business Entity Search & New-Business Monitor — Secretary of State (KYB API)

A business entity search and KYB API across official US Secretary of State business registries. Search registered companies by name, or monitor for newly formed LLCs and corporations on a schedule for fresh B2B leads. Bulk and date-range capable, sourced from state open-data portals — no anti-bot, no key.

Six states, live-verified on every run: Colorado, Connecticut, New York, Oregon, Pennsylvania, Delaware — 14.3 million records in total.

Who it's for

• Lead-gen & RevOps teams sourcing newly registered businesses to sell to.
• KYB / KYC / compliance teams verifying that an entity is registered.
• Sales intelligence & data providers enriching company records.
• AI agents needing a "look up / monitor US companies" tool.

What it does

Two modes. Search returns registered entities matching a name across the selected states. New businesses returns entities registered since a date (or in the last N days), so a scheduled run becomes a recurring new-business alert; turn on Delta mode and each run emits only entities no previous run of the same query has seen.

It tells you when a state did not answer

This is the part most multi-source actors get wrong. Every requested state is health-checked live before any result is returned — row count, freshness, required columns, a closed-vocabulary check, and a negative control (a filter that must match nothing; if it matches something, the portal is ignoring the WHERE clause and every result would be an unfiltered slice of the registry).

A state that fails is excluded and named, on every single row:

• partial_results — true when at least one requested state did not answer
• states_ok / states_failed / states_failed_detail

So "no match in Oregon" and "Oregon never answered" are never the same answer. If no state can answer, the run fails loudly and bills nothing rather than returning a green, empty dataset that reads like a verified negative.

Example input

{
  "mode": "new_businesses",
  "states": ["CO", "NY", "PA"],
  "daysBack": 30,
  "maxPerState": 100
}

Search for a company instead:

{ "mode": "search", "states": ["CO", "CT", "NY"], "query": "Acme", "maxPerState": 50 }

Output fields

Every field below is declared in the dataset schema and appears on the results table. Fields are nullable: null means the state does not publish it, never "we checked and there is none".

Entity

Field	What it is
name	Legal business entity name.
entity_id	State-assigned entity / account / licence / filing number.
type	Entity type as the state codes it (DLLC, DOMESTIC LIMITED LIABILITY COMPANY, …).
status	Registration status or standing.
status_basis	Where status came from: source_column (a real status field — CO, CT), dataset_scope (implied by the dataset being active-only, so never independently verified — NY, OR, DE) or not_published (PA).
formation_date	The state's own date for the record — see date_basis.
date_basis	What formation_date actually is per state: entity_formation_date (CO), registration_date (CT), initial_dos_filing_date (NY), registry_date (OR), creation_date (PA), license_valid_from (DE). Delaware publishes a licence roster, not a corporate registry.
registration_date	CT: the state-filed registration date.
record_created_at	CT: the open-data portal's ETL load timestamp. Published separately so it can never be mistaken for a business date.
jurisdiction	Jurisdiction of formation.
state	Two-letter code of the registry that produced the row.
registry	Human-readable registry name.
trade_name	Trade / DBA name (DE).
license_valid_to	DE: expiry of the current business licence.

Registered agent

Field	What it is
registered_agent	The agent-shaped name this actor has always returned for the state.
registered_agent_name	The authoritative registered agent as the state labels it — CO agent columns including middle name and suffix, OR entity-of-record or agent person columns, PA full party name, NY registered_agent_name. null where the state files none.
registered_agent_source	Which upstream column the agent came from. For New York this reads ny_dos_process_name, because NY publishes a service-of-process contact — frequently the entity itself — rather than a registered agent. Read it before treating NY's value as an agent.
registered_agent_address	Agent mailing address where the state publishes one (CO, NY, OR).
registered_agent_role	PA: role of the named party on the filing (Organizer, Officer, …).
dos_process_name	NY: the service-of-process contact on file with NY DOS.

Address & geography

Field	What it is
address	Composed address for the record. On Oregon, read record_type first.
address_line2	Second address line (OR, NY).
city · zip · country	Components, where published.
county · county_code	County (NY, PA) and PA's numeric code.
latitude · longitude	State-published geocode (PA, DE). null where the state publishes none.

Oregon record shape

Oregon publishes one row per associated name, not per business: 1,568,968 rows over 562,802 distinct businesses. The address on a REGISTERED AGENT row is the agent's, not the company's.

Field	What it is
record_type	PRINCIPAL PLACE OF BUSINESS, REGISTERED AGENT, MAILING ADDRESS or AUTHORIZED REPRESENTATIVE. Filter on it to de-duplicate.
entity_of_record_name · entity_of_record_id	The corporate entity named on the record (e.g. a commercial registered-agent firm).
record_url	Oregon's deep link to this exact registry record.
record_id	CT: the portal's own row identifier.
source_url	Link to the official registry record or, where the state publishes no per-record URL, its registry search page.

Connecticut ownership flags (state self-certification)

woman_owned, veteran_owned, minority_owned, lgbtq_owned, disability_owned. null on every other state — those registries publish no such column, so the answer is "not checked", not "no".

Run integrity — on every row

Field	What it is
state_status	Outcome for the registry that produced this row. Always ok: a state that failed contributes no rows and is named in states_failed instead.
partial_results	true when at least one requested state did not answer. Computed from the actual per-state outcomes every run — never a hard-coded false.
states_requested · states_ok · states_failed · states_failed_detail	Exactly which registries answered, which did not, and why.
result_truncated	true when this state returned exactly maxPerState records — more matched than were returned. Socrata publishes no truncation flag, so this is the only signal.
state_rows_fetched	Records this state returned before delta filtering.
run_mode · delta_mode · delta_baseline_size	The mode, whether delta filtering ran, and how many entity IDs the cross-run baseline held beforehand (0 on the first run of a new monitoring scope — that first run is a full dump by design).
data_quality_flags	Per-record notes: future_dated (the state published a date in the future — Delaware has 41 licences dated up to 2092-07-11), no_geocode, no_registration_date_published, no_registered_agent_on_file. null when the record is clean.

Delta / monitoring mode

Set monitor: true and schedule the run. The baseline lives in a named key-value store and is scoped to the mode, state set and query, so a narrow run cannot overwrite the baseline a wide run built. The first run of any new scope emits everything (delta_baseline_size: 0); every run after that emits only entities it has not reported before. Records with no state-assigned ID are always emitted rather than silently dropped.

Use as an MCP tool

Exposed to AI agents via mcp.apify.com as a "search US business registries" tool — clean inputs (states, query, date window) and structured entity output with explicit per-source status, so an agent can tell a verified negative from an unavailable source.

FAQ

Which states are covered? Colorado, Connecticut, New York, Oregon, Pennsylvania and Delaware, from each state's official open-data portal.

Is this a KYB API? It is a screening tool, not a determination. It tells you what the state's own published extract says. For a legal opinion on standing, pull the certificate from the state.

Why does Delaware look different? Delaware does not publish its corporate registry as open data. The DE source is the active business-licence roster, so formation_date is a licence start date — date_basis says so on every row.

Can I get only new companies? Yes — mode: "new_businesses" with daysBack or sinceDate, plus monitor: true on a schedule.

Why did a state return nothing? Check states_failed and states_failed_detail on any row. If the state is not listed there, it answered and genuinely had no match. PA republishes roughly monthly, so a 7-day window can legitimately be empty — widen daysBack.

Any anti-bot or captchas? No — this reads open-data APIs, not lookup pages.

Source

State Secretary of State (and, for DE, Division of Revenue) open-data portals on Socrata: data.colorado.gov, data.ct.gov, data.ny.gov, data.oregon.gov, data.pa.gov, data.delaware.gov. Keyless. An optional free socrataAppToken raises rate limits.

Pricing (Pay Per Result)

Billed per entity record returned, with graduated discounts on paid Apify plans. A run that returns nothing costs nothing, and a run that cannot answer fails instead of billing.

Related actors

• KYB Company Verifier — company verification across national registries.
• License Verifier — professional/occupational licence lookup, 19 boards.
• GLEIF Ownership Graph — legal-entity identifiers and parent ownership.
• NYC Landlord Registry Lead List · ACRIS Deed Transfer Intel — entity-level property and owner intelligence that pairs with a registry lookup.