Read
The compose box and the recipient fields, on the device only.
Service
An AI privacy guard that catches identifiable client data and exposed recipient lists before the message leaves — and stays quiet when the message is clean.
Before a message is sent, the guard looks for what identifies a person: a national ID number (with a full check-digit verification, so it does not cry wolf), a date of birth, a labelled record number, an email address, a phone number, or clinical content sitting next to any of them. It also counts who can see each other, because the most common way a client list leaks is a To or Cc that should have been a Bcc.
You can always send. The guard names what it saw and shows the placeholder to use instead, without repeating the identifier itself. It runs on your device, counts recipients and discards them, and has no capability to make a network call — which is a property of how it is built rather than a promise about how it is used.
The compose box and the recipient fields, on the device only.
Check-digit validation, pattern detection, and a recipient-visibility count.
Names what it saw, never repeats it, never blocks the send.
What this page does not claim: This demo is real and runs entirely in your browser: the NRIC/FIN check-digit validation and the To/Cc exposure count are the same logic the product uses. Nothing you type leaves the page.