For healthcare & life sciences teams

HIPAA-Ready PII & PHI Redaction That Runs Inside Your Cloud

Stop sending PHI to third-party redaction APIs. Philter runs entirely inside your VPC, automates much of HIPAA Safe Harbor across all 18 protected identifier categories, and ships with purpose-built clinical NLP models - not a generic LLM and a prayer.

All 18 HIPAA Safe Harbor identifier categories, with a ready-made policy you tune and validate
Purpose-built clinical NLP - not a generic spaCy recognizer
Runs in your VPC: no BAA renegotiation, no PHI leaves your perimeter
Open source & auditable - permissive Apache license, every line on GitHub

Permissive Apache license Runs in your VPC No third-party API HIPAA / GDPR / CCPA ready Redacting PII and PHI since 2017

Built on open source already running across

Clinical notes pipelines
EHR data exports
Medical chatbots
Healthcare RAG / AI
Claims & billing systems
Research data prep

Try it live

Try it out! Select one of the industries and click Redact to redact the text.

Input

Patient Margaret Collins, born on 04/12/1978, with SSN 523-88-4021 was admitted to the ER at St. Luke's Medical Center. Her primary care physician, Dr. Howard Banks, can be reached at hbanks@stlukesmed.org or (555) 342-9187.

Redacted output

The redacted text appears here after you click Redact.

Do not enter PHI or PII.

Why healthcare teams pick Philter

HIPAA Safe Harbor, end-to-end

All 18 protected identifier categories under 45 CFR § 164.514 - from names and SSNs to dates, biometric IDs, and full-face photos - addressed by a single configurable pipeline you validate against your data. See the blueprint.

Clinical models, not generic NER

Specialized NLP lenses trained on clinical text - recognizes medical terminology, drug names, and clinical context, not just emails and phone numbers. Why this matters.

BAA stays clean

PHI never leaves your VPC - no third-party API endpoints, no multi-tenant service, no new vendor BAA to sign. Your existing AWS / GCP / Azure BAA still covers everything.

Audit-ready by design

Stateless processing, immutable policy versioning, structured audit logs, and a benchmarking harness (Philter Scope) that scores every policy change against precision and recall on gold-standard datasets.

Sized for clinical-pipeline volume

Bulk EHR exports, real-time clinical NLP at the message bus, or HL7 / FHIR streams - same engine, horizontally scalable. Deployment patterns cover the common shapes.

Open source - review the code, not the marketing

Your security and clinical-informatics teams can audit the actual logic, fork it, or run it air-gapped - no closed-source ML, no “trust our HIPAA black box.”

What we’ll cover on the call

Your current pipeline: where PHI flows, what’s already redacted, what’s been deferred.
HIPAA Safe Harbor coverage: a real gap analysis against all 18 identifiers, not a generic checklist.
Deployment shape: VPC architecture, BAA implications, integration points (HL7, FHIR, CSV exports, S3, Kafka).
Cost picture: per-instance economics vs. per-character SaaS APIs at your actual volume. Most teams cross the break-even sooner than they expect.
A concrete next step: marketplace deploy, embedded engineering, or a no-go - whichever fits.

No NDA needed for the conversation. We won’t ask you to send sample data - we’ll walk you through how Philter would handle the patterns you describe.

See how Philter handles your clinical text

Bring a representative sample of the data your pipeline actually sees. We'll show you exactly what Philter catches, what it misses, and what to tune - on a 30-minute call, on your terms. Detection is probabilistic, so measure precision and recall on your own data before production; because Philter runs in your VPC, you remain the data controller for the output.

Or deploy Philter yourself →