Paste any AI response containing tokens. PrivacyScrubber swaps them back to originals — locally, instantly, with zero server contact.
ChatGPT responds using your tokenized text with [NAME_1], [EMAIL_1]...
Copy the AI response into the Reverse Scrub zone.
[NAME_1] → "John Doe", [EMAIL_1] → "john@company.com" — instantly.
![Zero-Trust PII Detection pipeline — raw document with PII passes through the PrivacyScrubber shield, producing anonymized tokens like [NAME_1] and [EMAIL_2]](/features/pii-detection-hero.png)
PrivacyScrubber's reverse scrub (un-mask) processes your data 100% locally in browser memory. No server ever sees your content — verified by our Airplane Mode test. This Zero-Trust Data Sanitization (ZTDS) architecture meets enterprise security standards out of the box.
A lawyer asks ChatGPT to draft a legal brief using scrubbed case facts. The AI output references [NAME_1] and [NAME_2]. Reverse Scrub restores the real party names for the final document.
Marketers scrub customer lists, use AI to generate personalized outreach, then un-mask tokens to produce ready-to-send emails with real names and company details.
HR uses AI to draft interview feedback notes. After AI generates structured assessments using tokenized candidate data, Reverse Scrub restores names for the final HR record.
Researchers use AI to summarize anonymized interview data. Reverse Scrub lets them restore participant identifiers for internal reports while keeping published versions anonymized.
During the initial scrub, PrivacyScrubber creates a session map: { "[NAME_1]": "John Doe", "[EMAIL_1]": "john@co.com" }. This map lives in browser RAM only, and is used to un-mask tokens.
The session map is permanently erased. This is by design — zero-trust means zero persistence. You'll need to re-scrub the original text to create a new map.
No. Each session creates a unique map. Tokens from a previous session cannot be un-masked because the map no longer exists anywhere.