Try a risk check

Your own IP is filled in. Put a BIN and a billing country next to it and see what Fraudex would say about the order. 3 free checks a day.

Why another fraud score?

Most fraud APIs hand you a number like "fraud_score: 87" and nothing else. When a real customer gets blocked, you can't tell them why, and you can't tell whether the tool is wrong. And the transaction scoring that combines card and connection data usually sits on plans that cost hundreds of dollars a month.

Fraudex does the opposite. Every point in the score comes from a named signal with a sentence explaining it, so you can show your support team exactly what happened, override a single rule, or set your own thresholds. It is included in the IP + BIN plan, with unlimited checks.

What it checks

  • The connection. Tor exits, VPNs, proxies, hosting and bot networks, from the same data as the VPN detection API.
  • The card. Brand, type, level, issuing country and whether it is prepaid, from the BIN.
  • Whether they agree. A card from one country, a billing address in a second and a connection from a third is the classic stolen-card pattern.
  • The email. Throwaway inbox services, if you send the address.
  • Card testing, live. Fraudex remembers your last 15 minutes of checkout attempts (BIN and IP only). One IP trying many different cards, or one BIN hit from many IPs, is flagged on the next attempt.

Card testing detection built in

Card testers run hundreds of stolen cards through a checkout in minutes, rotating IPs so each attempt looks new. A per-request fraud score can't see that, because each request looks fine on its own. Fraudex sees the pattern across your attempts: the fourth different card from the same IP gets a many_cards_one_ip reason, the sixth a card_testing_burst, and a BIN hammered from many IPs a bin_attack. Read more in how to spot and stop card testing.

Add it in one line

Call /api/risk before you charge the card, on every attempt, including ones that end up declined. Block on block, add 3-D Secure or an email code on review, and let allow through. Every signal and its points are listed in the docs.

Questions

Do I send the full card number?

No. Only the first 6 to 8 digits (the BIN). That identifies the bank and card type, never the cardholder, so it keeps you out of most card-data compliance scope.

How much does it cost?

A risk check uses one IP lookup and one BIN lookup. The IP + BIN plan ($70 for 30 days) and Complete plan ($99) make both unlimited. The Free plan includes 50 checks a month. See pricing.

Can I change the thresholds?

Yes. Use the score or the individual reasons codes and ignore decision. Many shops review everything over 30 and only block on Tor or card testing.

What data do you keep?

For card-testing detection, the BIN and IP of each attempt are kept in memory for 15 minutes, per API key, and then dropped. Nothing else from the request is stored beyond the usage log of your own API calls.