US Business Registry Scraper — New Companies by State

Get the companies that registered in a US state last week — with name, entity type, address and registered agent — before anyone sells them anything.

This Actor reads business registries published as open data by US states and returns one clean, normalised row per company, whatever each state calls its own columns. Six states are wired in today, covering 8.4 million entity records the Actor can reach plus New York's daily filings feed, and the underlying feeds refresh daily.

No API key, no login, no proxies. It queries the states' own open-data endpoints.

What you get

Every state names its columns differently. This Actor maps them all onto one shape:

• companyName, tradeName, registryId — the identifier the state assigned
• entityType — normalised where the state publishes a code instead of a label (Texas CL becomes Limited Liability Company). The value you get back is accepted as entityType input and always returns rows of that type; because the match is a contains one, a label that is contained in another label brings that one too (asking Texas for Limited Partnership also returns Foreign Limited Partnership).
• status — Good Standing, Delinquent, Dissolved, and each state's own wording
• registrationDate — the field that makes this a lead source rather than a directory
• address, city, postalCode, country, county, jurisdiction (where the entity was originally formed)
• addressType — which address the row carries, because states do not publish the same one: principal address (CO), business address (CT, DE), principal place of business (OR), taxpayer address (TX), address designated for service of process (NY)
• agentName and agentAddress — the registered agent, where the state publishes it as a person. Colorado is the only wired state with agent columns and it names a person in 2,203,866 of its 3,097,268 rows; the 710,453 rows whose agent is a company (agentorganizationname) come back with agentName: null
• filedBy — who submitted the filing, where the state publishes it (New York)
• ownershipFlags — certifications the state records: woman-owned, veteran-owned, minority-owned, LGBTQI-owned
• validUntil for states that publish licences rather than incorporations
• sourceDataset, sourceUrl, source — full provenance, and raw on request

Coverage today

State	Dataset	Records	Note
Colorado	Business Entities	3,097,268	Full registry, status and registered agent, updated daily
Texas	Active Franchise Taxpayers	3,415,044	Organisational codes decoded; 194,522 rows (5.7%) carry a non-active right-to-transact code and activeOnly drops them
Oregon	Active Businesses	559,186 reachable of 564,236	1,573,153 published rows, one per associated name; we read only the 559,194 principal-place-of-business rows, so the 5,050 companies that publish no such row cannot be returned — see Honest limits
Connecticut	Business Registry Master	1,292,899	Carries the ownership-certification flags
New York	Daily Corporate Filings	24,524 formations	A rolling 30-day feed of all 95,549 filings; the default keeps only formation and authority filings
Delaware	Business Licences	67,189	Licences with validity dates, not incorporations

Who uses this

• Sales and lead generation — every company formed in your state last week, filtered by name or city. New businesses buy accountants, insurance, software, equipment and services, and they buy early.
• Recruiters and agencies — spot new entities in a sector before they appear anywhere else.
• Competitive and market research — count formations by city, type and month to see where an industry is growing.
• Compliance and diligence — confirm an entity exists, in which state, since when, and who its registered agent is.
• Developers and AI agents — one normalised schema across six registries, callable over the API or as an MCP tool.

Example output

A real row, from a run of { "states": ["CO"], "registeredAfter": "2026-08-19" } on 2026-08-21 (null fields omitted):

{
  "companyName": "Biting Bullets LLC",
  "registryId": "20268035642",
  "state": "CO",
  "stateName": "Colorado",
  "entityType": "Limited Liability Company",
  "status": "Good Standing",
  "registrationDate": "2026-08-19",
  "address": "8772 W Arbor Ave",
  "city": "Littleton",
  "postalCode": "80123",
  "country": "US",
  "addressType": "principal address",
  "jurisdiction": "CO",
  "agentName": "QUINTON KELLY",
  "agentAddress": "8772 W Arbor Ave",
  "sourceDataset": "data.colorado.gov/4ykn-tg5h",
  "sourceUrl": "https://data.colorado.gov/d/4ykn-tg5h",
  "source": "US state open-data portals (Socrata)"
}

Input

Field	What it does
states	Two-letter codes: CO, TX, OR, CT, NY, DE. Empty = all six. An unsupported code is an error — the run never silently falls back to querying every state.
query	Company name contains — solar, dental, logistics.
registeredAfter	Only entities registered on/after this date, YYYY-MM-DD. This is the lead filter. A date the calendar does not have (2026-13-45) fails the run with a message instead of quietly returning nothing. What the date means differs by state — see Honest limits.
registeredBefore	Upper bound on the registration date, same validation. A window whose end precedes its start is rejected rather than run.
city	City of the address that state publishes. Every row reports which address that is in addressType.
entityType	Legal form, e.g. Limited Liability Company, Nonprofit, Corporation. Matched as contains, so corp finds every corporation type — and, like any contains match, a short fragment also hits inside a longer word (AB matches 330,577 Oregon rows through LiABility), so use a whole word. Colorado and Texas publish codes: a term that IS one of their published codes (DLLC, CL, and also the undocumented ones such as CO FO or TX AB) matches that code exactly, and anything else is matched against the code labels — never against the code text, so a Colorado Limited Partnership query returns 43,890 rows and no longer the 58,377 that included DLLP/FLLP limited liability partnerships.
activeOnly	Drop entities the state reports as not active. Default true.
newRegistrationsOnly	New York only: keep formation and authority filings, drop biennial statements and dissolutions. Default true.
includeRaw	Attach the untouched state record as well.
maxItems	Hard cap on rows exported (and charged) across all states. 0 = unlimited.

Every company registered in Colorado this month:

{ "states": ["CO"], "registeredAfter": "2026-08-01" }

Solar businesses across five states:

{ "states": ["CO", "TX", "OR", "CT", "NY"], "query": "solar", "activeOnly": true }

New Denver companies only, with the raw state record attached:

{ "states": ["CO"], "city": "Denver", "registeredAfter": "2026-08-01", "includeRaw": true }

Companies actually formed in New York this month — formations only, no biennial statements:

{ "states": ["NY"], "registeredAfter": "2026-08-01", "newRegistrationsOnly": true }

Run it on a schedule

Save your filters as a Task, set registeredAfter to a recent date, and schedule it daily or weekly. Attach an Apify integration and the new registrations land in Google Sheets, Slack, an e-mail, a webhook, or Make and Zapier for your CRM.

Deduplicate on registryId together with state — registry IDs are unique per state, not nationally.

Use it from code or from an AI agent

Start runs and read the dataset over the Apify API or CLI, in JSON, CSV, Excel or XML. The Actor is also exposed as an MCP tool through the Apify MCP server, so an agent can ask for "LLCs registered in Denver this month" in a single tool call and get structured rows back.

Pricing

Pay per event: a small fee for each company row exported. There is no browser and no proxy — only API calls — so platform usage on top is negligible. Use maxItems for a hard ceiling.

Where the data comes from

Each state publishes its own registry on a public open-data portal (Socrata). This Actor queries those endpoints directly and normalises the results; sourceDataset and sourceUrl on every row point back to the exact dataset.

Honest limits. Coverage is six states, not fifty — states differ widely in what they publish as open data, and several publish nothing. The datasets refresh on the state's own schedule (daily for most of these), so this is a fast feed, not a real-time one. Officer and shareholder details are not part of these datasets. This Actor is independent and not affiliated with any state agency.

What each state can and cannot answer, because the differences are large enough to change what a filter means:

• New York publishes a rolling 30-day feed of filings, not an entity register. Only 24,524 of its 95,549 rows are formations; the rest are biennial statements of long-established companies, name changes and dissolutions, and a re-filing carries the date of the re-filing, not of the incorporation. newRegistrationsOnly (on by default) restricts the feed to formation and authority filings — leave it on if you want new companies. The dataset also has no principal-address column: address/city come from the address the entity designates for service of process, which is frequently its registered agent's, so NY city results skew heavily towards Albany (15,582 of 95,549 rows). county is the entity's own principal-office county and is the more reliable NY location field.
• Oregon publishes one row per associated name — principal place of business, registered agent, mailing address, authorised representative — so its 1,573,153 rows are 564,236 companies. We read only the 559,194 principal-place rows, which is why a city filter returns the company's own city and not its agent's, and why the Oregon count above is not 1.57 million. That costs recall, and here is the size of it: 5,050 of the 564,236 Oregon companies publish no principal-place-of-business row at all — mostly limited partnerships, e.g. PACIFIC VIEW GARDENS LIMITED PARTNERSHIP (registry 10014181), which has only a mailing-address and a registered-agent row. Those 5,050 are returned by no filter, not even a plain name query, so 559,186 of Oregon's 564,236 companies are reachable and the run log says so on every Oregon query. Oregon also publishes no status column at all — its dataset is active registrations by construction — so activeOnly is a documented no-op there rather than a filter that silently did nothing.
• Texas covers franchise taxpayers, not every filed entity, and its registrationDate is the date the taxpayer became liable for franchise tax — close to the formation date, but not the same field. The dataset is published as "active", yet 194,522 rows (5.7%) carry a right-to-transact code other than A; activeOnly removes them.
• Delaware publishes business licences, not incorporations: rows carry licence validity dates instead of formation dates, and entityType is a licence activity category (GENERAL SERVICES, RESIDENT CONTRACTOR) rather than a legal form, so a legal-form term matches nothing there. No Delaware licence in the dataset is currently expired, so activeOnly changes nothing for it.
• Connecticut publishes ownership certifications, a billing city and a business type (LLC 833,746, Stock 343,980, Non-Stock 65,181, Limited Partnership 19,031, LLP 4,029) — the words Corporation and Nonprofit are translated to Stock/Non-Stock before querying. Limited Partnership returns the 19,031 limited partnerships only; the 4,029 LLP rows are a different legal form and are reached with LLP.

Related Actors

Same engine, same official-source approach:

• AusTender (Australia) — federal contracts, winners and expiry dates
• UK Public Tenders — Contracts Finder and Find a Tender in one run
• Swiss Public Tenders (simap.ch) — tenders and awards with prices, bid counts and weighted criteria
• Swiss Commercial Register (SHAB) — new companies, changes, deletions and bankruptcies
• EU Tenders (TED) — European procurement notices and contract awards
• France Company Register — SIREN/SIRET, executives, official labels
• GLEIF LEI — LEI lookup, lapsed-LEI screening, parent companies
• Brazil CNPJ — Receita Federal company data and partners

---

Italiano

Estrae le imprese registrate negli stati americani dai portali open data ufficiali: ragione sociale, identificativo, forma giuridica, stato, data di registrazione, indirizzo e agente registrato. Sei stati, 8,4 milioni di record societari, aggiornamento quotidiano. Il filtro sulla data di registrazione lo rende una fonte di contatti commerciali freschi: le imprese appena costituite comprano servizi, software e assicurazioni subito.

Deutsch

Exportiert US-Firmenregisterdaten aus offiziellen Open-Data-Portalen: Firmenname, Registernummer, Rechtsform, Status, Registrierungsdatum, Adresse und Zustellungsbevollmächtigter. Sechs Bundesstaaten, 8,4 Mio. Datensätze, tägliche Aktualisierung. Über das Registrierungsdatum filtern Sie neu gegründete Unternehmen als frische B2B-Leads.

Français

Exporte les registres d'entreprises des États américains depuis les portails open data officiels : raison sociale, identifiant, forme juridique, statut, date d'immatriculation, adresse et agent enregistré. Six États, 8,4 millions d'enregistrements, mise à jour quotidienne. Le filtre sur la date d'immatriculation fournit des prospects B2B récents.