About this project
We built this because the first step of brand work is finding out whether a name is already taken, and that first step is usually harder than it should be.
Who is behind it
This site is built and run by one independent developer, based in Shanghai, China. There is no company behind it, no law firm on retainer, no editorial team and no advertising. On a topic this close to legal advice, that is a limitation worth stating plainly rather than dressing up.
Being small is also why there is no author byline on these pages. We would rather leave the byline empty than invent a persona with credentials we do not have. What we offer instead is the opposite of marketing: the method is written out in full, the data source and its gaps are listed, the numbers on this site are the ones our own pipeline printed, and corrections are read by a person.
TODO: Replace with a real, verifiable description of the team/company. Google explicitly treats fabricated creator profiles (fake names, AI headshots, invented credentials) as deceptive. Leave this as TODO until it is true.
Why we built it
USPTO publishes its register as large annual XML volumes. That is excellent for researchers and awkward for a founder who just wants to type in a name. We did the unglamorous work of downloading and parsing the full archive, auditing every status code, and compressing it into small edge-side shards that answer a name lookup in a few milliseconds.
The search is free to try. We would rather someone discover a conflict before spending money on packaging, domains and a filing than find out after.
What we did ourselves
The entire pipeline was built and measured here: 91 annual volumes (11.50 GiB) plus 307 daily incremental files downloaded and byte-checked, 13,993,768 case-file entries parsed from the annual volumes and the daily deltas merged on top (9,036,198 updates, 429,675 new filings, 304,358 cancellations), 144 distinct status codes audited one by one, and 49,158 compressed objects (244.7 MiB) published to edge storage. Query cost was measured on the deployed infrastructure against the previous index build: 5.7 ms of CPU for an exact lookup and 11.8 ms for a fuzzy lookup.
- Data parsing, status classification and index construction were done in-house.
- Every status code classification has a documented source; see the status-code reference.
- Measured numbers come from our own benchmarks, not from vendor claims.
Contact and corrections
If a record looks wrong, tell us. We publish a corrections channel and an editorial policy because accuracy matters more on a legal-adjacent topic than on almost anything else.
Write to the address on the contact page. If you are disputing a record, include the USPTO serial number and the TSDR link; that is usually enough to settle it in one exchange.