USPTO Trademark Applications & Status Feed

Clean, flat US trademark records from the official USPTO daily trademark data. Get one useful row per mark with its word mark, serial and registration number, owner and address, goods and services, international class, status and dates, and run a changes-only mode that returns just the marks that moved since your last run.

You pay only for a delivered trademark record. Filtered-out rows, unchanged rows in changes mode, and malformed records cost nothing. There is no start fee.

No API key needed

The USPTO bulk data API requires a key. The actor supplies one, so there is nothing to sign up for and nothing to paste. Start a run and you get rows.

If you would rather run under your own USPTO account and its own quota, request a key at https://data.uspto.gov/apikey and paste it into the apiKey field. It is stored as a secret and is never written to the dataset or the log.

What a record is

One flat row per trademark case file:

• identity: serialNumber, registrationNumber (null until it registers), isRegistered
• the mark: markIdentification (the word mark), markDrawingCode, markTypes (trademark, service mark, collective, certification)
• status and timeline: statusCode, statusDate, filingDate, registrationDate, abandonmentDate, publicationDate, renewalDate, transactionDate
• owner: ownerName, ownerCity, ownerState, ownerCountry, plus a full owners list with street, postcode, nationality and legal entity code for marks with more than one owner
• classification: internationalClasses, primaryClass, and a classifications list carrying the US classes and the first-use dates per class
• what it covers: goodsAndServices, the actual recited goods and services text
• proceedings and maintenance: oppositionPending, cancellationPending, renewalFiled, section8Filed, section8Accepted, section15Filed, intentToUse, supplementalRegister, foreignPriority, and more
• provenance: sourceUrl, retrievedAt, confidence, jurisdiction, dataSource, and in changes mode changeType

Every value appears verbatim in the USPTO record, and sourceUrl is that mark's own public TSDR page, so any row can be checked at the source. Dates are normalised to ISO (YYYY-MM-DD). USPTO writes an absent date as 00000000 and an absent registration number as 0000000; both are reported as null rather than as a real value. Nothing is guessed.

Sample output

A real slice of one day's file (abbreviated):

markIdentification	serialNumber	registrationNumber	statusCode	filingDate	ownerName	ownerState	internationalClasses
BTF	50013146		630	2026-07-25	Bridge to Turkiye	NC	036

A pending application has no registration number. That is reported as null, not as zero, so you can filter on it.

Two modes

• full: every matching record in the window.
• changes: only marks that are new to you, or whose status, registration, abandonment, renewal or owner has moved since this account's last run. The state is tracked per account, so it is your timeline. The first changes run has no baseline, so it returns everything matching once, then tracks deltas.

Filters

• lookbackDays: how many days of USPTO daily files to read, counting back from today. USPTO publishes on business days, so 3 typically covers the last two or three published files.
• markContains: keep only marks whose word mark contains this text.
• ownerContains: keep only records where an owner's name contains this text. This is the portfolio watch.
• internationalClasses: Nice classes, for example 009 for software, 025 for clothing, 042 for tech services. A leading zero is optional.
• ownerStates: two-letter US state codes.
• statusCodes: USPTO status codes, for example 630 for a new application awaiting examination or 700 for registered.
• registeredOnly, liveOnly: drop pending applications, or drop marks carrying an abandonment date.
• maxRecords: cap on delivered and billed records. Defaults to 1000, because one USPTO day holds around 9,500 records. Set 0 for no cap. In changes mode nothing is lost to the cap; the rest arrives on a later run and is never billed twice.

What to expect

• USPTO publishes one file per business day, and each holds every application or registration that had activity that day. Measured on the 2026-08-01 file: 9,457 records. A busy day is several times that.
• A daily file is large, 9 MB to 49 MB compressed. The actor streams it rather than loading it, so a wide lookback is a matter of time rather than memory, and a run that reaches its time limit stops cleanly and delivers what it fetched.
• The raw bulk data is public and free directly from USPTO with a key of your own. What this actor sells is the flat typed record, the owner and class and goods-and-services extraction out of a large XML document, the filtering, and the per-account change feed, so you do not have to download and parse hundreds of megabytes of XML to find the handful of marks you care about.

Source and scope

What this collects: public US federal trademark registry records, covering marks and the businesses that own them. Source: the official USPTO Open Data Portal bulk product "Trademark Full Text XML Data (No Images) - Daily Applications", accessed over the documented API with a declared identifying User-Agent. No scraping, no access bypass. Scope: United States federal trademark records. Every row is stamped jurisdiction=US. Billing: charged only on a delivered trademark record; rejects are free; no start fee. Provenance: every row carries its TSDR source URL and the retrieval time.

These are public business records. This is a data tool, not legal advice, and a trademark record is a public filing rather than an opinion on anyone's rights.

See also

More clean, pay-only-for-results data tools from Pono Data:

• USPTO Patent Search & Changes Feed - US patent records from the USPTO Open Data Portal
• Federal Contract Opportunities Feed - US federal contract opportunities from SAM.gov

Full catalog: https://apify.com/thoob