Traditional scam protection relies on blocklists: someone reports a bad number, a human reviews it, and eventually it gets blocked. By then, the scammer has moved to a new number. Static lists are always weeks behind.
Scambook works differently. Our database of phone numbers, email addresses, and scam scripts is built to learn continuously — every single user interaction makes it smarter.
How the network learns
When you check a message, our AI doesn't just give you a verdict. It extracts anonymized scam signals: the sender's number or email, the structure of the script, the payment method demanded, the urgency tactics used. Those signals join millions of others in the database.
That means when a scammer switches from one phone number to another but keeps the same script, we still catch them — the script itself is fingerprinted. And when they invent a new script but reuse an email address, we catch that too.
The network is built on AI trained on the internet's collective record of scams — millions of publicly reported scam messages, senders, and scripts — and it keeps sharpening every time someone runs a check.
Privacy is the foundation
None of this requires storing your messages. Message contents are analyzed and discarded; only the anonymized scam indicators are kept. Your grandmother's texts to you never enter any database — but the scammer's script that someone else received does.
It's a simple flywheel: more checks → smarter database → better protection → more people protected. Every user makes every other user safer.
