NYC UCC Liens & Federal Tax Liens - ACRIS Personal Property

Every recorded UCC financing statement and federal tax lien filed in New York City: 4,552,980 documents from the city's ACRIS personal-property register, joined to their 11M party rows so a filing carries its debtor and secured-party names.

This is the lien and KYB surface - the thing a credit, onboarding or vendor-risk check reaches for. Not real property (that is the separate ACRIS Real Property dataset).

Low cost

From $0.00005 per record, down to $0.00003 at Gold. Pay only for the records you actually receive - the whole register stays queryable, you are never billed for the scan behind it.

Screen a company for liens

mode=debtor returns every filing naming a debtor (or any party), each with the document's type, date, borough and amount:

{ "mode": "debtor", "debtorName": "ACME", "maxResults": 50 }

Set exactName=true to match the whole name ("ACME CORP"); by default it matches a name prefix, which is what the source's index is fast on. Narrow to a role with partyRole (1 = Debtor, 2 = Secured Party, 3 = Third Party).

What is in a filing

• Type - decoded from the source's code: INIT / INIC initial UCC1, CONT continuation, TERM termination, ASGN assignment, AMND amendment, RLSE release, FL an IRS federal tax lien, FTL other federal lien, RFL release of federal lien, plus the amendment / withdrawal / partial-release and bankruptcy variants (29 codes in all).
• Document amount - the lien amount where stated. Note most UCC filings carry 0 (a UCC1 states no amount); federal tax liens carry the figure.
• goodThroughDate - the UCC continuation deadline, i.e. whether the filing is still live.
• Borough - decoded (1 = Manhattan 5 = Staten Island).
• Collateral - decoded (Fixture Filing, Cooperative, Manufactured Housing ).

Search documents

• Type - docType="FL" (IRS liens), "TERM" (terminations).
• Where - borough (1-5).
• When - dateFrom / dateTo (ISO).
• Amount - minAmount / maxAmount.
• Collateral - collateral (F / C / A ).

{ "docType": "FL", "dateFrom": "2026-01-01", "borough": "1", "maxResults": 50 }

Look up one document

mode=document returns the document plus every party on it (name, role, address) - a crfn is accepted as well as a documentId:

{ "mode": "document", "crfn": "2026000219331" }

Or roll it up

mode=aggregate returns one count row per group - document type, borough, collateral type or recording year. Every group comes back; this mode is not cut off by maxResults.

INIT    1,008,416
TERM      990,506
FL        377,567
CONT      248,045
ASGN      235,488

Example output

One debtor's filing - mode=debtor:

{
 "platform": "nyc-acris",
 "documentId": "2026080400059001",
 "crfn": "2026000219331",
 "docType": "TERM",
 "docTypeDescription": "UCC3 TERMINATION",
 "recordedDatetime": "2026-08-04",
 "borough": "Manhattan",
 "documentAmount": "0",
 "collateralDescription": "FIXTURE FILING",
 "partyRole": "Debtor",
 "partyName": "DEMETRIUS, JANE",
 "partyCity": "QUEENS",
 "partyState": "NY",
 "partyZip": "11429"
}

Notes on the data, from the source

• One row = one recorded document (or, in document / debtor mode, one (document, party) pair). document_id is essentially unique on the master.
• The debtor / secured-party names are in the parties table, not the document row - this actor joins them, which is what makes it a screening tool rather than a bare document list.
• document_amt is 0 on most UCC rows - a UCC financing statement states no amount. Do not read the column as a transaction value.
• recorded_borough and doc_type are codes in the source; both are decoded on output and shown as readable text alongside the code.
• Name search is prefix-based by default. The parties table is 11M rows and a leading-wildcard contains-match is a full scan the source times out on; prefix and exact matches are index-fast. Use exactName=true for a whole-name match.
• The register is refreshed daily, not in real time - ACRIS is a recording-once archive, so each record's recordedDatetime is its own filing date and the newest recordedDatetime in a result set is how current the data is.