Extracting data from trademark registration certificates
Each national trademark office designs its own registration certificate, and the USPTO's layout has changed over the decades. IP counsel for an acquirer or a lender review every registration in a brand owner's portfolio. Certificates arrive as downloads from the office's records or as scans in a data room.
Diligence needs the registration number, the owner, the register, each international class with its goods and the registration date. The IP schedule records them against each mark. A docketing system calculates the Section 8 and renewal deadlines from the registration date.
What makes trademark registration certificates hard
The mark itself is artwork
A design mark prints as an image, so its wording and design elements come from the artwork and the description paragraph that accompanies it.
Classes bundle many goods
One registration can cover several international classes, each with its own goods list and first use dates, so values have to stay with their class.
Paper and electronic certificates mix
Older registrations survive as scans of paper certificates, while newer ones are electronic, so one portfolio mixes ornate scanned pages with born-digital files.
Parse, extract, verify
Parse
Parse turns the document into reading-order Markdown and a hierarchy of blocks (text, tables, figures), and grounds each block to its exact location on the page.
Extract
You define the fields in a JSON schema, and Extract returns their values in a consistent, structured format, each linked back to where it appears in the Markdown.
Verify
Each value links back to where it appears, down to the character range in the Markdown, and the page and bounding box in the source document, so every result can be verified.
One trademark registration certificate, end to end
About this document: A USPTO registration certificate for the AMERICAN NINJA WARRIOR shield design mark, owned by Tokyo Broadcasting System Television, two portrait pages. Page 1 is the certificate: a bold left column of identifiers beside prose on the owner and the goods, with the mark and the seal as images. Page 2 sets out the filings that keep the registration alive. Page 1 is featured.
Processing objective: The objective is to record the registration's identity, the details an IP schedule lists for every mark a company owns. The identifiers sit in a column of their own beside the prose, and the serial number shares a line with the application's filing date.
Demonstration fields: Registration number, International class, Register, Owner, Serial number
The document
2 of 2 pages
Page 1shown below
Page 2
Where each value came from
registration_number
{
"value": "5,886,375",
"page": 1,
"range": {
"end": 117,
"start": 98
},
"box": {
"xmin": 0.05529,
"ymin": 0.38199,
"xmax": 0.25822,
"ymax": 0.40256
}
}international_classes[0]
{
"value": 21,
"page": 1,
"range": {
"end": 157,
"start": 144
},
"box": {
"xmin": 0.05572,
"ymin": 0.43936,
"xmax": 0.1759,
"ymax": 0.45695
}
}register
{
"value": "Principal Register",
"page": 1,
"range": {
"end": 188,
"start": 169
},
"box": {
"xmin": 0.05598,
"ymin": 0.49652,
"xmax": 0.25452,
"ymax": 0.51591
}
}owner.name
{
"value": "Tokyo Broadcasting System Television, Inc.",
"page": 1,
"range": {
"end": 253,
"start": 190
},
"box": {
"xmin": 0.33266,
"ymin": 0.38533,
"xmax": 0.80822,
"ymax": 0.39843
}
}serial_number
{
"value": "87-981,735",
"page": 1,
"range": {
"end": 993,
"start": 955
},
"box": {
"xmin": 0.33341,
"ymin": 0.61626,
"xmax": 0.61025,
"ymax": 0.62831
}
}How to try it yourself
Everything on this page is reproducible. The sample, its schema, and the Parse and Extract API outputs are published. You can repeat the same two calls from whichever surface you already build on.
- Claude Skilldescribe the job; the agent writes it
- CLIno code at all
- MCPexpose ADE docs as an agent tool
- REST APIany language
- Pythonlandingai-ade
- TypeScriptlandingai-ade
Documents into data
Accurate
The best accuracy on the difficult documents other tools get wrong, such as messy scans, complicated tables, checkboxes and handwriting.
Auditable
Every answer cites its sources. Values come back with an audit trail (page, bounding box, character range) so a human can review them.
Agentic
Applies multi-step reasoning, and plans before it acts. Handles documents it has never seen before and adapts to variations.
API-first
Composable APIs that drop into any tech stack and can be called from anywhere. Easy to build on and embed in your pipelines and products.
Affordable
Cost-competitive rates with complete price transparency. Spend controls before you run, and every job reports its exact cost.
Enterprise-grade
Built for production and scale from day one. Fully observable and monitored, running securely in our cloud or yours.
Built for high-volume, high-throughput production
- Zero data retention
- Priority rate limits
- Enhanced support
- VPC and on-prem deployment
Available on Team and Enterprise plans. VPC and on-prem deployment require an Enterprise plan.
