USPTO Trademark Search API: Bulk Name Availability Check

Check hundreds of brand names, product names or slogans against official USPTO trademark records in one run. For every name you get a conflict risk flag (high / medium / low / none), the reason, and the best matching US trademarks: mark text, serial and registration number, live/dead status, filing and registration dates, owner, Nice classes, a goods/services snippet and the match type (exact, sound-alike or contains).

Built for founders naming a startup, Amazon/Etsy/print-on-demand sellers screening designs and listing titles, agencies shortlisting brand names, domain investors, and IP teams who want a fast first pass before a full clearance search.

📅 USPTO data current through: 2026-11-01 (refreshed weekly from the USPTO's official daily trademark files; every result also carries dataThrough).

Not legal advice. This is an automated screening of USPTO records, not a clearance opinion. A "none" or "low" result does not mean a name is free to use or register (common-law rights, state marks, foreign marks, design marks and likelihood-of-confusion analysis are out of scope). Talk to a trademark attorney before you file or launch.

How it works

1. The Actor ships with a search index of the full USPTO trademark database (about 12.9 million US word marks, 4.5 million of them live, from 1870 to today), built from the USPTO's official bulk data on the USPTO Open Data Portal (Trademark Full Text XML Data, annual backfile + daily files). The date the data runs through is in every result as dataThrough.
2. Each name is normalized (case, spaces, punctuation and accents ignored) and matched three ways:
   • exact: same letters and digits (Nike = NIKE = N.I.K.E)
   • phonetic: sounds alike (Nyke, Nikki for Nike), using a Metaphone sound key plus a Jaro-Winkler similarity check of at least 0.75 to keep noise down
   • contains: the mark contains every word of your name (NIKE AIR, NIKELAB)
3. Results are filtered by your Nice classes and live/dead status, ranked (exact > phonetic > contains, live first, your classes first) and a risk flag is set.

No scraping of USPTO websites and no third-party data: only USPTO's documented bulk data products.

Risk flag

Level	Meaning
high	A live mark with exactly the same name exists in one of your classes (or in any class if you set none)
medium	A live sound-alike mark in your classes, or an exact live mark only in other classes
low	Only marks that contain your words, or only dead marks
none	No match found in the index

The flag always looks at live marks, even if you ask for dead marks only.

Input

Field	Default	What it does
terms	–	Names / phrases to check, one per entry (up to 10,000 per run)
serialNumbers	–	Optional USPTO serial numbers to look up directly (8 digits)
registrationNumbers	–	Optional USPTO registration numbers to look up directly
niceClasses	all	Nice classes 1-45, e.g. ["25"] clothing, ["9","42"] software
status	live	live, dead or all
matchTypes	all three	Any of exact, phonetic, contains
maxResultsPerTerm	20	Matches listed per term (counts always cover all matches)
includeGoodsServices	true	Add a goods/services snippet for live marks
usptoApiKey	–	Optional: your own free USPTO Open Data Portal key to also apply the newest daily USPTO files in your run

Example: screen three candidate names for a clothing brand.

{
  "terms": ["Nike", "Spellkin", "Zorblaxitron"],
  "niceClasses": ["25"],
  "status": "live",
  "maxResultsPerTerm": 10
}

Output

One dataset item per checked name or number. Real example (trimmed to one match):

{
  "inputType": "term",
  "term": "Nike",
  "normalizedTerm": "nike",
  "phoneticKey": "NK",
  "riskLevel": "high",
  "riskReason": "Exact live mark 'NIKE' (serial 98029846) in class 25",
  "exactLiveMatch": true,
  "matchCount": 91,
  "matchCountCapped": false,
  "exactLiveCount": 8,
  "exactDeadCount": 0,
  "phoneticLiveCount": 54,
  "phoneticDeadCount": 0,
  "containsLiveCount": 29,
  "containsDeadCount": 0,
  "exactLiveInOtherClasses": 25,
  "matches": [
    {
      "matchType": "exact",
      "similarity": 1.0,
      "mark": "NIKE",
      "serialNumber": "98029846",
      "registrationNumber": "7401694",
      "live": true,
      "statusCode": 700,
      "status": "Registered",
      "statusDate": "2024-05-28",
      "filingDate": "2023-06-06",
      "registrationDate": "2024-05-28",
      "owner": "Nike, Inc.",
      "niceClasses": [
        "10",
        "18",
        "21",
        "25",
        "26",
        "27"
      ],
      "goodsServices": "Hair accessories, namely, hair ties and hair bands; Yoga mats; bags specially adapted for yoga mats; Lunch bags not of paper, lunch boxes; water bottles sold empty; Sports bags, handbags, backpacks, duffle bags, purses, shoulder bags, tote bags, gym...",
      "pseudoMark": null,
      "tsdrUrl": "https://tsdr.uspto.gov/#caseNumber=98029846&caseType=SERIAL_NO&searchType=statusSearch"
    }
  ],
  "niceClassesFilter": [
    "25"
  ],
  "statusFilter": "live",
  "dataThrough": "2026-11-01",
  "checkedAt": "2026-11-02T21:40:00+00:00",
  "disclaimer": "Automated screening of USPTO records, not legal advice and not a clearance opinion. Consult a trademark attorney before filing or launching.",
  "error": null
}

Main fields:

• riskLevel, riskReason, exactLiveMatch
• matchCount plus counts per type and status (exactLiveCount, phoneticLiveCount, containsLiveCount, exactDeadCount, …) and exactLiveInOtherClasses
• matches[]: mark, matchType, similarity, serialNumber, registrationNumber, live, status, statusCode, statusDate, filingDate, registrationDate, owner, niceClasses, goodsServices, pseudoMark, tsdrUrl (opens the official USPTO TSDR record)
• dataThrough: date of the newest USPTO data used

Invalid inputs (blank, symbols only, malformed numbers) come back as items with an error and are not charged. A run summary is saved as RUN_SUMMARY in the key-value store. The dataset has two views: Risk overview (one row per name) and Matching marks (one row per match).

Pricing

Pay per event, no subscription:

Event	Price
Term checked (each unique name, serial or registration number, including names with zero matches)	$0.01
Actor start	$0.00005
Invalid inputs and duplicates	free

100 names cost about $1.00; 1,000 names about $10.00. Use the run's maximum cost setting to cap spend; the Actor stops cleanly when it is reached.

FAQ

How fresh is the data? The bundled index is refreshed every week from the USPTO's official daily trademark files (new filings, status changes such as live to dead, owner changes). The date is at the top of this page and in every result as dataThrough. For same-week freshness, add your own free USPTO API key (usptoApiKey) and the run applies the newest daily files first (adds roughly 10-60 seconds, no extra charge).

Does it check design (logo) marks? It checks the words of every mark, including the word element of design marks. It does not compare images.

Why is a dead mark relevant? Dead marks are not enforceable registrations, but they show past use and can hint at common-law rights. They are hidden by default (status: live).

What about state, EU or other foreign trademarks? Only US federal (USPTO) records are covered.

Is a name with risk "none" safe? No. It means no similar word mark was found in USPTO records by this automated search. A trademark attorney can run a full clearance search.

Is my USPTO API key stored? No. It is a secret input used only inside your own run to download USPTO daily files, and it is never written to the dataset, logs or key-value store.

Where does the data come from? USPTO Trademark Full Text XML Data (annual backfile and daily applications files), downloaded from the USPTO Open Data Portal. Source: USPTO. This Actor is not affiliated with or endorsed by the USPTO.