Philter

Note

This project is currently in experimental and not mean to be used in production right now.

Philter is an application for PII/PHI redaction and anonymization. It provides endpoints to filter sensitive information from text, explain identified spans, and manage redaction policies and contexts.

Contexts support referential integrity and can be used to store and retrieve redactions across multiple requests.

For the Java version of Philter, see philter-java.

Features

• PII/PHI Redaction: Identify and redact sensitive information such as SSNs from text.
• Explainable Redaction: Get detailed information about why specific spans were identified.
• Policy Management: Create, list, and delete redaction policies.
• Context Storage: Store and retrieve redactions based on context names, supporting both In-Memory and MongoDB backends.
• Prometheus Metrics: Built-in metrics for health, token counts, redaction counts, and context counts.
• Secure API: Optional Bearer token authentication and HTTPS support with self-signed certificates. (Recommended to use your own certificates in production.)

Quick Start

Using Docker Compose

The easiest way to run Philter with its MongoDB backend is using docker-compose:

docker-compose up --build

This will start the Philter API on https://localhost:8443.

Using Makefile

You can also build and run the binary directly:

make build
./go-philter

API Usage Examples

Refer to the examples.sh script for common API operations using curl.

For more detailed technical information, see DEVELOPER.md.

License

Copyright 2026 Philterd, LLC. "Philter" is a registered trademark of Philterd, LLC.

This project is licensed under the Apache License, Version 2.0. See the LICENSE file for details.