Fraud score: device, IP and bots

Score a sign-up or order from IP reputation, device velocity, bot signals and order details.

Score a sign-up, listing or order before you accept it. Send what you know about the session: the IP address, a device fingerprint from your client, the shopper's browser details and the order. You get a 0 to 100 score (higher is safer), a risk tier and a suggested action.

Send the check#

curl -sS -X POST 'https://api.trustverify.co.uk/api/v1/fraud/score' \
  -H "Authorization: Bearer $TRUSTVERIFY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "email": "buyer@example.com",
  "ip": "203.0.113.10",
  "deviceFP": "fp_3c9a1e7b2d4f",
  "customerId": "cus_demo_001",
  "orderValue": 120,
  "orderCount": 0,
  "paymentMethod": "card",
  "billingAddress": {
    "country": "GB",
    "postcode": "SW1A 1AA"
  },
  "shippingAddress": {
    "country": "GB",
    "postcode": "SW1A 1AA"
  }
}'

At least one of email, ip or deviceFP is required, otherwise you get 400 validation_failed.

What is checked#

Signal	Input	What it does
IP reputation	`ip`	VPN, Tor and IP risk lower the score. If no IP result is available, signals.ip.available is false.
Device velocity	`deviceFP`	Counts how many of your customers have used the same device. Live keys only: test keys never read or write it.
Bot signals	`userAgent`, `botSignals`	The shopper's browser details that you forward. We do not score your server's own User-Agent.
Address mismatch	`billingAddress`, `shippingAddress`	A different country or postcode lowers the score.
Order	`orderValue`, `orderCount`, `paymentMethod`	High values, no order history and prepaid or gift cards lower the score.

Read the result#

action is approve, review or block. riskTier is low, medium, high or critical.
scoreSource is signals when an IP result was used, and heuristic otherwise. A heuristic score never returns approve.
Each entry in signals says whether you provided it, so you can see what the score was based on.

Reference: POST /api/v1/fraud/score.