Scan and decide on every transaction in milliseconds.

MeraFraud scores fraud risk in real time by combining payment behavior and user signals with a trained machine learning model — approve, flag for review, or block.

Transactions (30d)

—

Total monitored transactions

Blocked

—

High risk, auto-rejected

Loss Prevented

—

Estimated, from blocked transactions

Model ROC-AUC

—

Avg response: —

Global Risk Map

Recent flagged transactions across your markets

Approve Review Block

Live Activity

Most recent scored transactions by location

Loading…

Test a Transaction

Build a transaction scenario and see MeraFraud's risk score and reasoning.

Transaction Amount (€)

Ratio to Average (×)

Account Age (days)

Customer Lifetime Value (€)

Time Since Last Transaction (min)

Transactions in Last 24h

Hour (0–23)

Items in Cart

Failed Payments / Login Attempts (last 7 days)

—

⚠ API unreachable — this result is a local JS approximation. Start the Flask API for real model scores (see README).

What Is the Model Looking At?

Global feature importances from the RandomForest model