Behavioral Analysis

The behavioral analysis engine builds risk profiles over time by tracking device fingerprints, IP addresses, geographic patterns, and transaction velocity.

Signal	Trigger	Score	What It Detects
Fraud Score	Internal ML score > 30	0–100	Aggregate behavioral anomaly score
Multiple Devices	> 3 devices on one account	Up to 70	Account takeover, credential sharing
Multiple IPs	> 5 IPs on one account	Up to 60	Proxy/VPN usage, bot activity
Multi-Country Activity	> 2 countries on one account	Up to 80	Impossible travel, geographic anomalies
Custom Flags	Engine-detected anomalies	By severity	Session hijacking, device spoofing, velocity spikes

For new senders with no transaction history, the engine returns "New User (No History)" with score 0. Behavioral profiles strengthen as more transactions are processed.

Tip: Include deviceId and ipAddress in your transaction requests to enable the full power of behavioral analysis.

Overview

Core Concepts

Features

Modules

API Reference

Behavioral Analysis