See exactly what happens to your prompt

Paste any text — PII is detected and sanitized before it ever reaches your LLM.

Input prompt140 chars

Sanitized output

Sanitized text appears here…

import requests

result = requests.post(
    "https://api.raipii.com/v1/sanitize",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={"text": your_prompt, "mode": "fake_substitute"}
).json()

safe_prompt = result["sanitized_text"]
session_id  = result["session_id"]

Regex + Comprehend

Dual-engine detection. AWS Comprehend for context-aware PERSON/ADDRESS, regex for SSN/credit cards/JWTs.

Session restore

After your LLM responds, call /restore with the session_id to swap fakes back to real values automatically.

HIPAA mode

Regex-only detection — Comprehend never sees your data. Toggle per customer record.