Technical & Engineering
Patent Document
Extract bibliographic data, classifications and claims from patent filings across offices and languages.
The Challenge
Patents are long, multi-language and inconsistently laid out across offices. Claims carry nested dependencies, classification codes follow several schemes at once, and the same family appears under different numbers per jurisdiction.
Industries
Automotive, chemicals, machinery, electronics, pharmaceuticals, IP law.
Who Uses It
R&D, IP and patent departments, competitive intelligence, patent attorneys.
BASE Schema
Publication number — Text
Application number — Text
Publication date — Date
Applicant / assignee — Key value set (name, street, postal code, city, country)
Inventors — Table
Title — Text
Abstract — Text
IPC / CPC classification — Table
Claims — Table
Cited references — Table
Recommended pipeline settings — Context (internal technology taxonomy for classification)
OPTIONAL Validation
Publication and application numbers checked against office formats; priority date precedes publication; claim numbering continuity verified; family members deduplicated.
Frequently ASked Questions
How can patent documents be turned into structured data?
Bibliographic data, classifications and claims are extracted into one schema regardless of issuing office or language, so filings become searchable and comparable.
Can multi-language patent portfolios be processed together?
Yes — documents in different languages map to the same target schema, so an entire family can be reviewed in one consistent structure.
See more use cases.
Get inspired by further sample data schemas.