Data sources
Where the numbers come from.
Every figure on TrademarkMind is traceable to a public source. We rely on three feeds, all from the U.S. Patent and Trademark Office, all freely available, all updated on USPTO's own publication schedule.
1. USPTO Trademark XML feed (TRTDXFAP)
The primary source. USPTO publishes a daily XML file documenting every change to every U.S. trademark application, registration, and post-registration filing. This includes:
- The mark itself (literal text, design code, image)
- Owner identity, address, type
- Attorney/correspondent of record
- Filing date, classes, goods/services description
- Status code and prosecution history events
- Logo image references
Public feed location: bulkdata.uspto.gov/data/trademark/dailyxml/applications/. USPTO publishes new files most weekday evenings (US Eastern); we typically have them ingested within ~3 hours of publication.
2. USPTO trademark image archive
Alongside the text records, USPTO publishes the trademark images themselves: the logos and design marks the office examined. These remain the property of their respective trademark owners; USPTO makes them publicly accessible as part of the trademark register. Our Logo Similarity Search tool searches across this image data.
3. Nice Classification reference (WIPO)
The Nice Classification system (the 45 international classes used to categorize goods and services in trademark filings) is maintained by the World Intellectual Property Organization (WIPO). We use the latest published edition for class names, descriptions, and goods/services examples. WIPO publishes annual updates; we apply them within a few weeks of release.
4. WIPO Madrid Monitor daily files (ST.96)
International registrations under the Madrid System come from WIPO's Madrid Monitor daily update files: the same ST.96 XML WIPO publishes each day, downloaded every morning and applied in date order. Each file carries the complete current state of every international registration that changed that day: holder, mark, Nice classes, the countries it designates, and the full history of decisions by each designated office (grants, provisional and final refusals, oppositions, invalidations). Bundled mark images are kept for the logo tools.
Because the feed is a daily delta, our copy holds every registration with activity since we began collecting (July 2025) rather than the complete historical backfile; dormant registrations older than that are not included. Currently 592,895 international registrations, through the October 1, 2026 file.
Attribution: International data from the World Intellectual Property Organization (WIPO) Madrid Monitor, re-used under the Creative Commons Attribution 4.0 licence per WIPO's open-access policy. We normalize statuses and outcomes; WIPO bears no responsibility for the integrity or accuracy of the data as presented here. Mark images remain the property of their holders.
What we deliberately exclude
- Other national registers. Apart from the WIPO Madrid data described above, national systems (and EU trade marks filed directly at EUIPO) are out of scope. Madrid Protocol filings that designate the U.S. also appear in the USPTO feed once USPTO examines them, so those are covered from both sides.
- State trademark registrations. Each U.S. state has its own (mostly unused) trademark registry; we do not track them. Federal registration is what our analyses are about.
- Common-law marks. Trademarks that are used in commerce but never federally registered have legal force, but no public record. They aren't in our data.
- Internal USPTO correspondence. We have access only to what USPTO makes public. Examiner reasoning beyond what appears in the public Office Action is not in our dataset.
Last refresh
The most recent successful sync from USPTO completed at . Numbers shown elsewhere on the site reflect data as of that timestamp.
Citing TrademarkMind
If you reference our analyses in your own work (academic, journalistic, professional), link to the specific page rather than the homepage, and include the date you accessed it. Our numbers update as USPTO data updates, so a fixed date stamps the version you cited.
For the underlying USPTO data, cite USPTO's own bulk feed (linked above), not us. We're the interpretation layer on top of their data, not the data itself.