Account takeover, caught at the signal level
Device fingerprint mismatches, credential reuse patterns, and session anomalies scored before the transaction reaches your authorization gateway.
Learn moreTransaction risk scoring API
Score every transaction in the 40ms before authorization
Riskgrove catches the account-takeover and card-testing patterns that rules-based engines quietly wave through. Built for fintech and marketplace risk teams.
Live transaction stream
Scoring
TXN ID
Amount
Risk score
Status
TXN-8472A91
$127.00
12
Cleared
TXN-8472B24
$1,299.00
87
Flagged
Pattern detected
4 distinct card numbers tested from the same device fingerprint in 8 minutes. Velocity pattern consistent with credential stuffing pre-validation.
TXN-8472B25
$89.50
23
Cleared
TXN-8472C01
$450.00
94
Blocked
Consumer lending platform
E-commerce marketplace
BNPL fintech
Neobank payments
Gig economy platform
Early-access pilot program participants
The problem
Rules engines only catch what someone already wrote a rule for
Attack patterns evolve faster than rule update cycles. By the time your team encodes a new pattern, attackers have already rotated to a variation your rules don't match.
- New account-takeover vectors take weeks to reach a rule definition
- Card testing attacks spread across cards to stay below per-card velocity thresholds
- Rules built for one fraud cohort generate false positives on legitimate customers
- Post-authorization fraud analysis arrives too late to affect the auth decision
txn.amount > 1000
match
card.country != ip.country
match
user.txn_count_1h > 5
no match
device.velocity_24h > 3
no match
card.test_pattern_score
no rule
Result: transaction authorized. $12,400 in chargebacks followed 6 days later.
How it works
One call. One score. Fits inside your existing authorization stack.
Ingest
Send transaction data to the API
POST transaction fields to the scoring endpoint before your authorization request completes. Fields: card token, device fingerprint, user ID, amount, merchant category.
Score
ML model evaluates 200+ signals
Riskgrove runs gradient-boosting and behavioral graph models against real-time and historical signals in under 40 milliseconds median latency.
Act
Receive score and top signal reasons
Response includes a 0-100 risk score, top contributing signal reasons, and a recommended action (approve / review / decline). Wire to your auth logic.
Capabilities
Three fraud patterns. One scoring layer.
Each detection module targets a distinct attack class with its own signal set and model architecture.
Card testing patterns, identified before the second attempt
Multi-card velocity scoring across device fingerprints, merchant contexts, and time windows that per-card rate limits can't see.
Learn moreVelocity context that goes beyond single-field checks
Multi-dimensional rate analysis across entity pairs (card-device, card-merchant, user-IP) surfaces coordinated attack patterns that single-field velocity ignores.
Learn moreEarly results
What risk teams in our pilot cohort found
"We had a card testing attack that cost us around $12K in chargebacks before a colleague pointed us to Riskgrove. The device-fingerprint velocity scoring caught the pattern in the first batch of transactions on the next rotation. That's a signal our rules engine just doesn't have."
48%
Reduction in false positive rate
Measured across our pilot cohort of 6 customers over 90 days. Range: 40-65% depending on transaction mix. Not a universal claim.
"We were seeing ATO on seller accounts every week and couldn't pin it with transaction-level rules alone. Riskgrove's behavioral graph signals flagged session patterns we had no visibility on. Four weeks in and ATO volume on seller accounts dropped to near zero."
<40ms
Median scoring latency
Measured at p50 in our production infrastructure. p99 is under 90ms. Measured from API receipt to response.
Integration
One API call. One risk score. Works with your existing stack.
Drop Riskgrove behind your payment gateway. No rule configuration, no ongoing tuning required.
POST /v1/score
// Request
const score = await fetch('https://api.riskgrove.com/v1/score', {
method: 'POST',
headers: { 'Authorization': `Bearer ${apiKey}` },
body: JSON.stringify({
card_token: 'tok_live_8f2a...',
user_id: 'usr_0019b',
amount_cents: 129900,
device_id: 'dvc_f7c3a9',
merchant_category: '5045'
})
});
// Response (38ms)
{
"risk_score": 87,
"recommendation": "review",
"top_signals": [
"device_multi_card_velocity",
"card_first_use_high_amount",
"session_anomaly_score"
],
"latency_ms": 38
}Pricing
Simple per-call pricing. No setup fees.
Ready to start?
Score your first transactions free
Book a 30-minute call with our team. We'll walk through your current fraud stack and show you where Riskgrove fits.