Philter® Lenses

Sensitive information in text presents challenges across many industries. Doctors, healthcare providers, and their business associates have patient information in medical records, lawyers and legal assistants have confidential client information in legal filings, and financial advisors and tax preparers have clients’ sensitive financial information in documents and communications, just to name a few.

Available Lenses

A lens’ primary responsibility is identifying persons’ names in text. Philter includes the General Lite Lens by default. This lens was created to provide good performance across many types of documents. The other Philter lenses available are:

  • General Lens – Created for general use to find and redact person’s names in text such as news articles, medical text, and other narrative text. This lens works best when your documents cover a range of types.
  • General Lite Lens – Same as the General Lens but optimized for efficiency.
  • COVID-19 Lens – Created specifically to find and redact person’s names in healthcare text regarding COVID-19.
  • Healthcare Lens – Created specifically to find and redact person’s names in general healthcare text.
  • Legal Lens – Created specifically to find and redact person’s names in legal text.

What is a Lens?

Philter’s Lenses are like camera lenses. If you are taking a picture of the Eiffel Tower you are going to get the best picture with a wide lens. Likewise, if you are taking a picture of a flower you will get the best picture with a short lens. Philter’s Lenses work the same way. The healthcare lens is optimized for finding and redacting sensitive information in medical text, the legal lens is optimized for legal text, and the financial lens is optimized for financial text. Using a lens that matches your type of text will provide improved filtering performance. The general lens is for cases in which your documents are not from any of those unique types.

Philter Lenses are created using modern natural language processing techniques. They are constructed using large amounts of hand curated text from each of the individual industries. This text is used to train natural language processing models optimized for each of the corresponding industries.

Get started with Philter!

Example Redacted Document

The document below was redacted using Philter’s Legal lens. In this example Philter redacted person’s names and dates.


Philter in the Cloud

Philter is available for immediate use on AWS, Microsoft Azure, and Google Cloud. Launch Philter in your cloud!