US Business Entity Search & New Business Filings

Run a US business entity search by company name, or pull the new business filings of the last days, straight from the open-data APIs that six states publish themselves: New York, Colorado, Connecticut, Oregon, Pennsylvania and Texas. It is built for compliance, KYB, credit and sales teams that need company records they can show to a reviewer: every row is company-level data from an official source, with no officer, owner or agent-person names in it. To try it, keep the sample input and press Start: 25 companies registered in New York and Colorado in the last 7 days, in a few seconds.

• Official sources only. Each row comes from a state's own open-data portal and says which dataset it came from, links back to the record, and carries the time the state last updated that dataset.
• Company data, no people. Name, register ID, entity type, status, formation date, jurisdiction, addresses and industry code. Officers, owners and agents who are people are never returned, and sole proprietors are skipped. You can check this in the dataset schema.
• Pay only for records found. $3.00 per 1,000 records ($2.00 on the Gold plan). A name that a register does not hold costs nothing and still gets an answer row.

What a record looks like

{
    "state": "NY",
    "entityId": "8039659",
    "name": "NEW FITNESS BOXING CONCEPT LLC",
    "entityType": "DOMESTIC LIMITED LIABILITY COMPANY",
    "status": "active",
    "statusRaw": null,
    "formationDate": "2026-10-06",
    "jurisdiction": "NY",
    "principalAddress": null,
    "mailingAddress": { "street": "418 BROADWAY, STE N", "city": "ALBANY", "state": "NY", "zip": "12207" },
    "county": "Albany",
    "registeredAgentName": "NORTHWEST REGISTERED AGENT LLC",
    "naicsCode": null,
    "sourceDataset": "Active Corporations: Beginning 1800 (New York State Department of State, data.ny.gov/n9v6-gdp6)",
    "sourceUrl": "https://data.ny.gov/resource/n9v6-gdp6.json?dos_id=8039659",
    "sourceUpdatedAt": "2026-10-07T11:26:00.000Z",
    "query": null,
    "scrapedAt": "2026-10-07T19:09:45.827Z",
    "found": true
}

How to search US business entities and new filings

1. Choose the states.
2. Either enter company names (one per line, each is searched in every chosen state), or set Registered within the last days to get a feed of new registrations. With neither, the run lists the register newest first; companies the register gives no date for come last.
3. Narrow it if you like: entity type, status, city, ZIP code.
4. Set Maximum records. You pay per record, so this is the cap on the cost of the run.
5. Run it. Results are in the dataset; download them as JSON, CSV or Excel, or read them through the Apify API.

A daily feed of new companies. Save the input as a task, switch on Only records not returned before and add a schedule. Each run then returns only the registrations that earlier runs of the task have not returned, so each one is delivered and charged once.

Name match. "Contains" finds the text anywhere in the registered name and puts the exact name first. "Exact name" compares the whole name, punctuation included ("Tesla, Inc." is not "Tesla Inc"). "Starts with" is the middle way and the safest choice for a lookup.

From an AI agent. The Actor can be used as a tool through Apify's MCP server (https://mcp.apify.com): the agent finds it by its name, enisbodlli/us-business-registry-search, and calls it with an input like the ones below. The input and dataset schemas describe every field, so the agent needs no further instructions. A run answers "is this company registered in these states" in a few seconds, and the row's sourceUrl lets the agent cite the state's own record.

The six states

State	Source dataset	Published by	Refreshed	Status	Agent	County	NAICS
NY	Active Corporations: Beginning 1800 (data.ny.gov, n9v6-gdp6)	Department of State	daily in practice	active entities only	organisations only	yes	no
CO	Business Entities in Colorado (data.colorado.gov, 4ykn-tg5h)	Department of State	daily	yes	organisations only	no	no
CT	Connecticut Business Registry, Business Master (data.ct.gov, n7gp-d28j)	Secretary of the State	daily	yes, with sub-status	no	no	yes
OR	Active Businesses - ALL (data.oregon.gov, tckn-sxa6)	Secretary of State, Corporation Division	weekly	active entities only	organisations only	no	no
PA	Registered Businesses in PA (data.pa.gov, xvd7-5r2c)	Department of State	monthly as stated	active registrations only	no	yes	no
TX	Active Franchise Taxpayers (data.texas.gov, 9cir-efmm)	Comptroller of Public Accounts	every few days	yes (right to transact business)	no	no	yes

Checked on 2026-10-07. sourceUpdatedAt on each row shows how fresh the state's file was when the row was read.

Pricing

Pay per event. You pay for results, not for compute time or storage.

Event	Free and Bronze	Silver	Gold
company-record: one company record found and stored	$3.00 per 1,000	$2.50 per 1,000	$2.00 per 1,000
Actor start	$0.00005 per run	$0.00005 per run	$0.00005 per run

Not charged: a name that a register does not hold (the row with found: false), sole proprietors that are skipped, records an earlier run already returned, and searches that fail.

Worked example. A task that pulls the last 7 days of new registrations in New York and Colorado every morning with Maximum records at 500 costs 500 x $0.003 + $0.00005 = $1.50 a run on the Free or Bronze plan, $1.25 on Silver and $1.00 on Gold. A single lookup of one company name in all six states that finds one record costs $0.003 + $0.00005 = $0.00305. The same lookup that finds nothing costs $0.00005.

The maximum charge you set for a run is respected: the Actor stops before the next record would pass it and says so in the status message. If a run fails or is aborted you pay only for the records that are already in the dataset. A run that is restarted or resurrected reads its own dataset first, so it never stores or charges the same record twice.

Input example

New registrations of LLCs and corporations in two states, delivered once each:

{
    "states": ["NY", "CO"],
    "registeredWithinDays": 7,
    "entityTypes": ["llc", "corporation"],
    "maxResults": 500,
    "onlyNew": true
}

Look up three names in all six states, at most five matches per state and name:

{
    "states": ["NY", "CO", "CT", "OR", "PA", "TX"],
    "nameQueries": ["Blue Bottle Coffee", "Northwind Traders", "Acme Holdings"],
    "matchMode": "startsWith",
    "status": "active",
    "maxResultsPerState": 5,
    "maxResults": 100
}

Output example

[
    {
        "state": "CO",
        "entityId": "20268245921",
        "name": "Commoban Inc",
        "entityType": "Domestic profit corporation",
        "status": "active",
        "statusRaw": "Good Standing",
        "formationDate": "2026-10-06",
        "jurisdiction": "CO",
        "principalAddress": { "street": "1500 N GRANT ST STE R", "city": "DENVER", "state": "CO", "zip": "80203" },
        "mailingAddress": null,
        "county": null,
        "registeredAgentName": "Registered Agents Inc",
        "naicsCode": null,
        "sourceDataset": "Business Entities in Colorado (Colorado Department of State, data.colorado.gov/4ykn-tg5h)",
        "sourceUrl": "https://data.colorado.gov/resource/4ykn-tg5h.json?entityid=20268245921",
        "sourceUpdatedAt": "2026-10-07T11:19:36.000Z",
        "query": null,
        "scrapedAt": "2026-10-07T19:09:45.827Z",
        "found": true
    },
    {
        "state": "TX",
        "entityId": "32108279921",
        "name": "ALFIE LABS INC.",
        "entityType": "FOREIGN PROFIT CORPORATION",
        "status": "active",
        "statusRaw": "Active",
        "formationDate": "2026-12-13",
        "jurisdiction": null,
        "principalAddress": null,
        "mailingAddress": { "street": "10601 CLARENCE DR STE 250", "city": "FRISCO", "state": "TX", "zip": "75033" },
        "county": null,
        "registeredAgentName": null,
        "naicsCode": null,
        "sourceDataset": "Active Franchise Taxpayers (Texas Comptroller of Public Accounts, data.texas.gov/9cir-efmm)",
        "sourceUrl": "https://data.texas.gov/resource/9cir-efmm.json?taxpayer_number=32108279921",
        "sourceUpdatedAt": "2026-10-03T08:40:57.000Z",
        "query": "Alfie Labs",
        "scrapedAt": "2026-10-07T19:09:46.102Z",
        "found": true
    },
    {
        "state": "OR",
        "entityId": null,
        "name": null,
        "entityType": null,
        "status": null,
        "statusRaw": null,
        "formationDate": null,
        "jurisdiction": null,
        "principalAddress": null,
        "mailingAddress": null,
        "county": null,
        "registeredAgentName": null,
        "naicsCode": null,
        "sourceDataset": "Active Businesses - ALL (Oregon Secretary of State, Corporation Division, data.oregon.gov/tckn-sxa6)",
        "sourceUrl": "https://data.oregon.gov/resource/tckn-sxa6.json",
        "sourceUpdatedAt": "2026-09-29T13:35:00.000Z",
        "query": "Alfie Labs",
        "scrapedAt": "2026-10-07T19:09:46.455Z",
        "found": false
    }
]

state and entityId together are unique within a run. Every field is always present; null means the register did not say. status is one of active, inactive and unknown; statusRaw keeps the register's own words. A formation date in the future is a filing with a delayed effective date. Rows from several states arrive in the order the portals answer, so sort by state if you need them grouped.

The run also writes a RUN_SUMMARY record to its key-value store: totals, what was skipped and why, and the outcome of every search.

Limits

Company-level data only. The dataset never holds the name, home address or ID of a natural person.

• Officers, directors, principals and owners are not returned. The Actor does not even request the columns that hold them.
• A registered agent is returned only when it is an organisation. An agent who is a person is left out.
• Sole proprietors, estates, partnerships of individuals and assumed business names with no organisation behind them are skipped and counted (soleTradersSkipped in the summary). In Oregon about four in five assumed business names are skipped for this reason.
• In New York the mailing address is the address for service of process. It is returned only when the name on it is the company itself or another organisation.
• Address lines are returned as the register writes them, with three exceptions that keep names of people out: a "c/o" or "attn" line, a first line with no number and no street word, and a line below the street that reads like a full name are dropped unless they name an organisation. A town repeated on a second line goes with the last rule; it is in city. A name that a filer wrote into the street line itself cannot be told apart.

What is not covered.

• Other states. Only states that publish their register through an official open-data API are read: the six above. The other registries, South Carolina among them, are websites that need a browser, a proxy or a CAPTCHA to read, and this Actor does not do that. Iowa and Hawaii no longer offer such an API (checked 2026-10-07).
• Detail pages and filing history: they are not in the open-data sets.
• UCC liens: a different source in each state, not included.
• Phone numbers, email addresses, contact checks and lead scores: not returned.
• A separate standing field: the datasets publish one status value, and statusRaw returns it unaltered.
• Lookup by entity number: not in this version. Search by name.
• A proxy setting or an API token input: there is none, because none is needed. The portals are read with plain requests, one per second each. A portal that throttles is asked again after a pause, and if it keeps refusing the run reports it; nothing is done to get around a refusal.
• Real-time register state: rows are as fresh as the state's published file. sourceUpdatedAt shows its age.
• City business-licence datasets: not included.

Per state.

• New York, Oregon and Pennsylvania publish active entities only. status is always active there and "status": "inactive" returns nothing from them.
• Texas rows come from the Comptroller's roll of franchise taxpayers, not from the Secretary of State's register. They have no agent, the address is the taxpayer's mailing address, the status is the Comptroller's "right to transact business", and the county is published only as a number, which the Actor does not return.
• Colorado type codes are written out from the Secretary of State's list. Six codes that the list does not explain (FO, GP, UNA, DLCA, CS, FLCA) are returned as published and count as "Other" in the entity-type filter. About 5,400 companies formed outside Colorado are filed under FO whatever their legal form (counted 2026-10-08), so an LLC or corporation filter leaves them out in Colorado.
• Pennsylvania does not say where a foreign registration was formed, so jurisdiction is PA for domestic ones and null for the rest.

Caps.

• Up to 500 company names and up to 1,000,000 records per run.
• Several states share "Maximum records" and the maximum charge of a run: each state gets an equal part first, and what one state does not use goes to the others. A state whose portal takes more than 30 seconds for its first answer can find part of its share taken. "Maximum records per state" sets a fixed cap per state instead.
• "Only records not returned before" remembers what a task returned for 400 days, up to 200,000 records per task and state. It stops a listing after 20 pages in a row with nothing new.
• One request per second per state portal, as their robots.txt asks. A name lookup takes one to two seconds per state, and a listing reads up to 1,000 records per request (200 in Oregon). The states are read side by side.
• Reading a whole register is slow at that pace: a million records take at least 17 minutes per state, and all of Oregon about an hour and a half. Set the run timeout to match. A run that times out keeps what it stored, and resurrecting it goes on from the page it had reached.

Colorado's notice. The State of Colorado asks applications that use its data to say the following, and it applies to every Colorado row: the data has been modified for use from its original source, the State of Colorado; the State makes no representation or warranty as to its completeness, accuracy, timeliness or content; the State disclaims all warranties, express or implied, including merchantability and fitness for a particular purpose; the data is subject to change as modifications and updates are completed; and the information is used at one's own risk. The State's own wording is at https://data.colorado.gov/terms. The other states publish their data as is, without warranty, too.

FAQ

Is it legal? The Actor reads public datasets that the states publish for reuse through a documented API, without logging in, and it returns company-level fields only. New York, Colorado and Connecticut state their terms on their portals; Colorado and Connecticut place the data in the public domain. You are responsible for how you use the data and for following the laws that apply to you. A company name can contain a person's name ("Jane Doe Consulting LLC"); that is the company's registered name and is returned as such.

A name lookup found nothing. You get a row with found: false for each state that does not hold the name, free of charge. Check the match mode first: "Exact name" needs the punctuation of the register. Then check the state: only the six above are read, and New York, Oregon and Pennsylvania list active entities only.

My scheduled run returned an empty dataset. With "Only records not returned before" that means nothing new was registered since the last run, or the state has not refreshed its file yet. Compare perState.<state>.sourceUpdatedAt in RUN_SUMMARY with the day before.

mailingAddress is null on many New York rows. Those companies named a person on the service-of-process line, and the Actor does not return a person's address.

A search failed. RUN_SUMMARY lists every search with its result and reason. "refused": the portal answered 401, 403 or 404, sent a challenge page, or throttled three searches in a row. "unavailable": it did not answer after five attempts (two when it timed out), or kept answering "too many requests". "changed": the state renamed a column of its dataset or sent something that is not the data; please open an issue on the Issues tab and it will be fixed. A portal that leaves three searches in a row without an answer gets no more requests in that run, and its remaining searches are marked the same way without being sent. Searches that worked are in the dataset either way. Resurrecting the run tries the "unavailable" searches again and charges only the records it adds; a new run of the same input returns and charges every record again.

Can a rerun charge me twice? Not within one run: a restarted or resurrected run skips every record that is already in its dataset. Two separate runs of the same search return the same records twice unless "Only records not returned before" is on.

More Actors from this developer

• North Data Scraper: German & European Companies
• Handelsregister Scraper: German Company Register
• European Company Registry Search
• Brazil CNPJ Scraper: Company Search & Lookup
• Company Jobs Search
• ATS Job Postings: Workday, Greenhouse, Lever & Ashby
• Workday Jobs Scraper
• Greenhouse Jobs Scraper
• Lever Jobs Scraper
• Ashby Jobs Scraper
• Website Contact Scraper
• Email Validator