Is the USPTO trademark API free?

Yes. The USPTO provides US trademark data free through TSDR, which returns per-mark records, and through bulk downloads of the full register. The cost is not the API; it is the engineering to turn raw, examiner-oriented data into a fast, forgiving search feature your users will accept.

Can I do a fuzzy trademark search with the USPTO API?

Not directly. The USPTO returns exact records for a known mark or serial number, but it has no built-in fuzzy or phonetic search, which is exactly what name clearance needs. You either build that matching layer yourself on top of the data, or use a source that already includes it.

Does the USPTO API show the attorney or firm behind a mark?

The attorney of record is in each record, but the public search will not let you query or aggregate by firm. To rank or list firms by filing volume you have to process the bulk register yourself, which is how the firm filing data on councl.io was built.

The USPTO Trademark API: What It Does and Where It Bites

The USPTO exposes US trademark data free through TSDR and bulk files, but with no fuzzy search, heavy formats and rate limits. It is a strong foundation, not a finished search feature.

The USPTO gives you US trademark data for free, which makes it sound like the job is done. It is not. The data is there, official and current, but it arrives in a shape built for trademark examiners and attorneys, not for the search box you want in your product. Knowing the gaps before you start saves a painful week.

Key takeaway: the USPTO is an excellent source and a poor finished product. It will tell you everything about a mark whose number you already have, and very little in the fuzzy, forgiving way a real user expects to search.

What you actually get

Two things, mainly. TSDR, the status and document system, returns the full record for a specific application or registration: status, owner, classes, filing and renewal dates, the prosecution history. And there is bulk data, the whole register as downloadable files, refreshed regularly. Between them you can see the state of any US mark and, with enough work, the entire register. The data is free and as authoritative as it gets.

Where it bites

The trouble starts the moment a real user types a name instead of a serial number. There is no clean fuzzy or phonetic search, the very thing clearance needs, so "is anything close to MOONGLO taken" is not a question the API answers for you. The formats are heavy and assume a machine that already knows the schema. Rate limits and the occasional outage are facts of life. And the bulk data is enormous, so standing up your own searchable copy is a project, not an afternoon.

The representative problem

Here is one that surprises people. If you want to know which law firm filed a mark, or to rank firms by how much they file, the public search does not group the data that way. The attorney of record sits in the record, but assembling every mark by a given firm means processing the whole register yourself. It is doable, we did it for the directory, but it is nobody's afternoon.

What it is good for, and not

The USPTO is the right foundation for anything serious about US marks. It is the wrong thing to hand a product team and expect a finished feature by Friday. The distance between the raw data and a search experience your users will accept, fast, forgiving, always up, is exactly the distance most teams underestimate.

We have already crossed that distance

The free trademark check on this site runs on USPTO data, with the fuzzy matching, caching and uptime handling that the raw API leaves to you. It is the proof, and you can use it right now.

If you are weighing the USPTO API for your own build, the fastest way to learn what it will really take is a short call, not a spike. See how we integrate trademark data and book a consult. You will leave knowing exactly which pieces are an afternoon and which are the iceberg.