Description

PatentEye is an API providing NLP based analytics of patents’ data. Its main features include:

• Topic modelling and extraction on patents’ abstracts
• Semantic search of patents’ abstracts and claims
• Organization and Author search for patents

Currently serving up-to-date data from the weekly United States Patent and Trademark Office (USPTO) releases.

Topic modelling and extraction on patents’ abstracts

By leveraging state-of-the-art Natural Language Processing (NLP) models, PatentEye is able to provide data on the topics describing patents’ submissions, allowing the user to understand which trends currently exist. This topics are made available as whole, characterizing the whole dataset of over 500k patents (from January 1st 2021 onward), but also in detailed fashion for each of the existing Organizations and International Patent Classification (IPC) Sections, Classes and Subclasses. You can also understand topic progression across time. Specifically, this API can help answer:

• What are the main topics characterizing current patent applications?
• How are these topics evolving on a weekly basis? Which are more frequent, which are losing traction and which are gaining momentum?
• Which topics are Organizations X and Y mostly working on?

These topic extraction and understanding capabilities can help users from areas as diverse as Market Research or Venture Capital quickly assess what are the main existing trends and use this data to inform their research and decisions.

Semantic search of patents’ abstracts and claims

Besides providing the user with the ability to do simple patent searches by keywords present in the abstract or claims of the existing patents, PatentEye also provides powerful semantic search once again leveraging state-of-the-art NLP models. Through this keyword and the more powerful query based search, the user can effectively:

• Search for patents containing specific keywords of interest (e.g., “call center; software”)
• Search for patents which are similar to a certain query of interest (e.g., “intelligent system for call center management”) at both the Abstract and Claims level.

These powerful features can help companies providing or dealing with Intellectual Property (IP) processes to efficiently validate the uniqueness of their own (or their clients’) inventions and claims.

Organization and Author search for patents

By collecting and organizing up-to-date information on existing patents, PatentEye also provides a quick way to assess which Organizations and Authors are patenting what. Examples of this data collection can be, but are not limited to:

• Keyword based search of Organizations and Authors
• Retrieval of patents published by a certain Organization or Author

Furthermore, besides helping answer these concrete questions and use cases, PatentEye can also serve as a stepping stone to any user or company wishing to create their own datasets and create their own analyses and models.

API documentation

Each of the existing endpoints is thoroughly described with usage and response examples.

Customization

Would you like to see some specific feature or analysis being provided? Do get in touch!

Future work

We are continuing to expand our offer so keep an eye out for new features. Data from other patent offices, such as the European Patent Office will be made available shortly