HIPAA de-identification workflows
HIPAA's safe-harbour method requires removing 18 specific identifier categories. Redactly covers the text-borne ones — names, geographic subdivisions, dates, contact details, record and plan numbers, device identifiers, and URLs — in a single call, so an extract can be shared or analysed without carrying them along.
Try it on the live API
No signup. 2,000 characters, 20 runs per 10 minutes.
Output appears here.
What this gives you
- Covers names, addresses, postal codes, dates of birth, phone, email, MRN, health plan IDs, device identifiers, IPs, and URLs.
- Health-specific types are separate, so you can redact clinical identifiers while keeping the clinical narrative.
- Deterministic output makes it practical to re-run the same extract and compare results.
- Safe harbour is a standard, not a product feature: validate the output against your own HIPAA review before relying on it.
Request
curl -X POST https://api.bitbanger.in/v1/redact \
-H "X-API-Key: rdct_your_key_here" \
-H "Content-Type: application/json" \
-d '{"text": "MRN: 004512378. Patient Jane Doe, DOB 1987-04-12, seen at 214 Marlborough Road, Bristol."}'Questions
Does this make my data HIPAA de-identified?
No product can guarantee that. Redactly removes the identifier categories it detects, which supports a safe-harbour review, but the determination remains yours.
Does it redact the clinical narrative?
Only if you ask it to. Conditions and medications are separate entity types, so you can choose to keep the clinical content.