Your text and documents never leave your network to be de-identified.

Your data contains sensitive information -- that's why you're here, right? Why would you want to upload or send your most sensitive information to a third-party system just for it to be redacted? Isn't that a security risk?

Our software runs in your cloud and never transmits your data anywhere outside of your cloud. Keep your sensitive information in your own cloud.

Learn about our no cost engagement process

Preventing Disclosure of Sensitive Information in AI-Generated Text

A policy layer is an important part of every AI-powered application because it inspects the AI-generated text to prevent sensitive information from being exposed. A policy layer can help remove information such as names, addresses, and telephone numbers from responses. Airlock was designed to be a powerful policy layer in your AI architecture.

Learn more about how Airlock can set guardrails for your AI applications.

De-identification Regulations

There are many industry regulations around de-identification and redaction of documents. Additionally, virtually every industry has sensitive information that needs protected even if there are not well-defined regulations for it.


The most well-known industry requiring redaction is healthcare due to HIPAA. Section 164.514(a) of the HIPAA Privacy Rule describes in detail how health information must be redacted.


The Federal Rule of Bankruptcy Procedure 9037 outlines how federal bankruptcy filings must be redacted. This rule applies to both electronic and paper documents.


PCI DSS and Gramm-Leach-Bliley Act (GLBA) have rules about how consumer information must be protected. Account numbers and other information must be safe-guarded.

Redaction of AI Training Data

Are you wanting to train or fine-tune an awesome new AI model but there is personal information in your training data? AI is a fast-growing industry currently without regulations like HIPAA to define how information should be de-identified, but that does not make de-identification any less important.

Redacting Files in Amazon S3 Buckets

Amazon Macie can find sensitive information in files in your Amazon S3 buckets. Philter can redact those files. It's a great combination!

Learn more about using Philter alongside Amazon Macie.

De-identify and Redact with Philter

Philter can identify and redact over 30 types of PII, PHI, and other types of sensitive information. It is highly configurable and gives you control over how PII and PHI should be redacted through features like consistent anonymization, redaction logic, and multiple filter profiles.

Get a free demo of Philter's capabilities