4️⃣ The mission
Detection is only half the problem
Moderators shouldn't have to watch chat 24/7
Human attention should build communities, not fight scams
Security shouldn't depend on who's online
🛡️ Protect communities without restricting users
3️⃣ The approach
Stop looking at messages the way scammers want you to.
Every detection rule is tested against real attacks and real conversations before it goes live.
If a rule catches legitimate discussion, it doesn't ship.
Every false positive becomes a permanent test case.
3️⃣ The approach
Stop looking at messages the way scammers want you to.
Every detection rule is tested against real attacks and real conversations before it goes live.
If a rule catches legitimate discussion, it doesn't ship.
Every false positive becomes a permanent test case.
2⃣ What's broken
Scammers don't post obvious scams. They hide behind markdown tricks, lookalike characters, redirects, shortlinks, and images.
A human can usually spot it.
Most bots can't.
2⃣ What's broken
Scammers don't post obvious scams. They hide behind markdown tricks, lookalike characters, redirects, shortlinks, and images.
A human can usually spot it.
Most bots can't.
1⃣ The problem
Years of running crypto communities. Hundreds of scams collected by hand. Dozens of moderation bots tested. The pattern was always the same:
‣ scammers got through
‣ users got restricted
‣ moderators cleaned up the mess
The scammers adapted
1⃣ The problem
Years of running crypto communities. Hundreds of scams collected by hand. Dozens of moderation bots tested. The pattern was always the same:
‣ scammers got through
‣ users got restricted
‣ moderators cleaned up the mess
The scammers adapted