Machine Learning
        for Patent Analysis

Search, verify and map the patent space.

See Examples      Run Free Trial

Proximity Reports

We leverage computational linguistics, machine learning, and high-dimensional visualization to analyze the patent landscape.

Advancements in cloud computing enable us to convert what used to be a labor-intensive and error-prone practice of searching for prior art into a far more accurate and faster report.

A search by an attorney can take weeks.
We deliver a comprehensive report in 10 minutes.

We calculate a high-dimensional Vector Space Model from a starting point of reference that you provide (such as an existing patent, a patent application, or the text of a white paper or article) and compare it to every other patent publication and patent grant at the USPTO. Most importantly, we cross-check every single word in the space to identify associations that are hard to find. While keyword searching can find the most obvious patents, finding the remaining 20% is extremely difficult.

At $500 per report, PatRF is affordable and comprehensive.

We import data from the USPTO several times a week so that our analysis is as up-to-date as possible. The latest patent number in our system is 9686895 and the latest publication number is 20170181364. You can't find a more up-to-date analysis.

See What PatRF Can Do For You

Run Free Trial

Who is using PatRF?

Our mission is to make the patent system more effective for all. As a research foundation, we do not differentiate between users - all parties receive equal access to the system. Our current service helps you to find prior art and assess the patent landscape. Future offerings will leverage Tensor Flow and AI methods to predict emerging activity in the patent space.


Investment Bankers


R&D Managers


prior art search

deal-making and due-diligence

cease & desist and litigation

freedom to operate and lay-of-the-land

strategic planning and portfolio tracking

Example Reports