Philter identifies, removes, and, optionally, masks Personally Identifiable Information (PII) and Protected Health Information (PHI) from natural language (unstructured) text. Philter is capable of identifying PII and PHI as described by the 16 HIPAA identifiers that pertain to textual information, such as phone numbers, email addresses, credit card numbers, social security numbers (SSNs), and patient names. Philter uses a combination of pattern matching, dictionaries, and natural language processing.

Through customizable filter profiles, Philter can be configured to identify and remove any combinations of PII and PHI from text. Conditional logic can be applied to each type of filter to allow for complete control over how the PII and PHI is manipulated, such as by truncating zip codes based on known census populations. Locations of identified PII and PHI can be persisted to a database for future reference.

Philter features an API for text filtering and allows for integration into your existing pipelines and systems. Philter's stateless behavior allows for load balancing and autoscaling. Open source client SDKs are available.