PII Detection & Redaction API
Enterprise-grade PII detection and anonymization, offline-first.
Detect and redact personally identifiable information across 50+ entity types and 12 languages. Runs entirely on your infrastructure — no data leaves your perimeter. Built on Microsoft Presidio with deterministic offline extraction for PDFs, images, and Office documents.
Featured endpoints
Extract + redact. Streams NDJSON: one ``extracted_page`` line per page, then a single ``redacted`` line, then ``done``.
Stream extracted text. NDJSON: one ``page`` line per page, then a ``summary``.
Start with a guide
Send your first redaction request in under 60 seconds.
API keys, bearer tokens, and rotation.
Built-in PII categories and custom entities.
Replace, redact, mask, and hash.
Consume NDJSON file-processing streams.
HTTP error codes, validation errors, retries.
Pricing
View full pricing- 1,000 redactions / month
- 5 MB max file size
- Community support
- 250,000 redactions / month
- 50 MB max file size
- Email support, 24h SLA
- 2,000,000 redactions / month
- 500 MB max file size
- Priority support, 4h SLA
- Unlimited volume
- Single-tenant or on-prem
- Dedicated capacity
