Detect. Anonymize. Restore.

Open-source PII detection with reversible vault tokenization. Runs 100% in your browser via WebAssembly. No data leaves your device.

WebAssembly 100% Client-Side MIT License 47 Recognizers

Try the Demo GitHub crates.io npm

47

PII Recognizers

18

Countries

1.2

MB WASM Bundle

0

Dependencies

Locale:

Input

Anonymized Output

Click "Detect PII" to analyze

The Vault generates deterministic tokens for each PII entity. The same input always produces the same token. Tokenization is perfectly reversible — every token maps back to its original value.

Original Text

Vault Tokenized

Click "Tokenize & Restore"

Restored from Tokens (lossless) every token perfectly reversed

Restoration appears here

Real-world use case: Send PII-safe text to an LLM, then restore PII in the response. This simulates the full anonymize → process → de-anonymize pipeline.

1. Input → 2. Anonymize → 3. Send to LLM → 4. Restore PII

Step 1 — Your Input (contains PII)

Summarize this patient record: Patient: John Doe, SSN 123-45-6789 Email: john.doe@hospital.com Diagnosis: Type 2 diabetes, prescribed Metformin 500mg. Insurance: Card 4532015112830366 Emergency contact: Jane Doe, +1-555-0199

Step 2 — Sent to LLM (PII removed)

Pipeline will show anonymized text here...

Step 3 — LLM Response (with tokens)

Simulated LLM response appears here...

Step 4 — Final Output (PII Restored)

The restored output with original PII will appear here...