USPTO TSDR API Documentation
Explore TSDR case identifiers, record retrieval and authorized access requirements.
Learn more →Data & Research
Trademark Database vs Trademark Search vs a Qualified Dataset
People often use “trademark database” to mean several different things: a search system used to discover marks, a record system used to inspect a case, or a structured dataset built for analysis or outreach. Those are different products.
Trademark research path
- 1Discover candidate marks
- 2Check case identifiers
- 3Verify with TSDR
Data company workflowNormalize / qualify / verify / deliver
On this page
Direct answer
Three different layers are often called a trademark database
USPTO Trademark Search is a discovery/search layer. TSDR is a record-level status and document layer. A professional commercial dataset is a third layer: a defined set of records normalized and qualified for a particular business question.
Compare the layers
| Layer | Primary purpose |
|---|---|
| Trademark Search | Find and filter federal trademark records. |
| TSDR | Inspect current status and documents for a known application/registration. |
| Raw export | Machine-readable or spreadsheet output from a source/search; may still need interpretation and cleanup. |
| Qualified dataset | Records selected under explicit rules, normalized, verified and structured for the target use case. |
Researching official case records with an API
When you already know a case identifier, the TSDR API supports programmatic status and document checks, separate from keyword discovery and commercial lead datasets.
Trademark Search vs TSDR
Choose the correct system for your particular research task.
Learn more →A bigger database is not automatically a better product
For commercial data, inclusion logic, freshness, provenance and field interpretation can matter more than the raw number of rows. A smaller correctly qualified cohort may be more useful than a massive unfiltered dump.
For commercial data, inclusion logic, freshness, provenance and field interpretation can matter more than the raw number of rows. A smaller correctly qualified cohort may be more useful than a massive unfiltered dump.
Additional context
A search system is optimized for discovery. A record system is optimized for case-level inspection. A commercial dataset is optimized for repeatable analysis, segmentation or operations. Confusing those layers creates bad expectations—for example, assuming that a search export is already deduplicated, that every visible email is a verified owner contact, or that one broad status flag is enough to define a prospect.
For AI and data workflows, a qualified dataset should make the semantics explicit: which source supplied each important field, which field drove inclusion, when the record was checked, and which transformations or exclusions were applied. That semantic layer is what turns a collection of rows into reusable intelligence.
Common questions
Is the USPTO trademark database the same thing as TSDR?
No. Trademark Search is the current federal search system; TSDR is the record-level status and document system.
Why not just export search results?
A raw export can be useful, but it usually does not encode the full business qualification rules, provenance, deduplication and field interpretation required for a professional use case.
Can AI systems use a structured trademark dataset?
Yes, subject to the customer’s lawful use and data-governance requirements. Explicit field definitions and provenance make downstream analysis more reliable.
Database access and data intelligence solve different problems
| Need | Best-fit layer |
|---|---|
| Find a trademark record | Use a discovery/search database such as the current USPTO Trademark Search system. |
| Verify a specific case | Use record-level status/document evidence such as TSDR. |
| Analyze a defined cohort | Use a structured dataset with explicit status/event, date and qualification fields. |
| Repeat the same commercial filter | Use a documented qualification profile and deduplication rule rather than a one-off manual search. |
| Audit why a row was included | Retain identifiers, provenance and verification metadata with the delivered data. |
Official U.S. sources
Need a custom trademark dataset?
Tell us the business question, not just “send the database.”