Fraudeny - Real-time fraud scoring for fintechs and platforms

Real-time fraud mitigation infrastructure

Fraudeny detects, scores, and blocks risky activity in real time through a single API. Built by a founder who has shipped fraud models in production. Goes live in hours, not months.

No credit card. No sales call. No 6-month contract.

POST /v1/score - Response
{
  "risk_score": 87,
  "risk_level": "HIGH",
  "explanation": [
    "IP country mismatch (0.35)",
    "Amount 4.2x above average (0.28)",
    "3 transactions in 5 min (0.22)",
    "First purchase in category (0.15)"
  ],
  "action": "BLOCK",
  "latency_ms": 47
}
LOWMEDIUMHIGH
47ms Scoring
Real-time risk scores with explainable AI for every transaction.
Production Experience
Built by a founder with experience shipping fraud and AI systems in production.
One API Call
POST /v1/score. Get risk score + explanation. That's it.
The Market Today

Fraud Detection Shouldn't Require a $500K Budget

Most companies either overpay for enterprise tools, waste months building in-house, or simply go unprotected.

Too Expensive
Enterprise fraud tools start at $50K/year. That's your entire engineering budget at Series A.
Too Slow
6-month implementation cycles. By the time you're live, you've already lost $200K to fraud.
Too Complex
You need one API endpoint, not a platform with 47 dashboards and a 6-month rollout.
4
Years Building Fraud Models
200ms
Scoring Latency
99.9%
Uptime SLA
95%+
Precision
590K+
Transactions Trained
$0
To Start
How It Works

From Zero to Protected in 3 Steps

No sales call. No 6-month contract. No procurement committee.

1. Get Your API Key
Sign up in 30 seconds. No sales call or contract required.
2. Send Transactions
POST your data. Get a risk score with explanation in 200ms.
3. Act on Intelligence
Block, review, or approve. Set up webhooks for alerts.

Fraudeny

Day 1: API KeyDay 1: First ScoreDay 2: Production

Enterprise Solutions

Month 1: Sales CallMonth 3: ContractMonth 6: Maybe Live
One API Call

Real-Time Fraud Scoring

import requests

response = requests.post(
    "https://api.fraudeny.com/v1/score",
    headers={"Authorization": "Bearer fd_live_..."},
    json={
        "amount": 499.99,
        "currency": "USD",
        "card_country": "US",
        "ip_country": "NG",
        "user_id": "user_456"
    }
)

score = response.json()
# Returns in 47ms

What You Get Back

Risk Score87/100 - HIGH
IP Country Mismatchweight: 0.35
Amount Anomalyweight: 0.28
Velocity Spikeweight: 0.22
New Categoryweight: 0.15

Every score includes SHAP-powered explanations. Know exactly why a transaction was flagged. Audit-ready for compliance.

Built for Real Fraud Operations

Intelligence You Can't Get Anywhere Else

Real-Time Scoring
Score any transaction in under 200ms. Ensemble model (XGBoost + IsolationForest + Autoencoder) runs in parallel.
Explainable AI
SHAP-powered explanations with every score. Know exactly WHY a transaction was flagged. Audit-ready for SOC 2 and PCI.
Extreme Imbalance Handling
Designed for rare-event fraud data, where the challenge is catching real loss without flooding the review queue.
Auto-Adapting Models
Models learn from your data without constant rule churn or manual retuning.
Ethical AI Built In
Bias detection and fairness monitoring built into the scoring workflow so teams can audit model behavior with confidence.
Webhook Alerts
Get notified instantly when high-risk transactions hit. Integrate with Slack, PagerDuty, or any webhook.
Where We're Heading

Beyond Transaction Scoring

Most fraud systems ask one question: "Is this transaction suspicious?" We think that's necessary but not enough.

Cybersecurity evolved from breach detection to attack surface management years ago. Fraud hasn't made that leap yet. We're building toward a world where fraud teams don't just react to suspicious events -- they map their entire fraud surface, understand what's exposed, and respond with auditable, deterministic actions.

The scoring API is our foundation. What comes next is the infrastructure to shift fraud prevention from event-centric to surface-centric.

Fraud Exposure Graph
Map every exploitable surface -- impersonation domains, payout flows, refund loopholes, onboarding gaps -- before attackers find them.
Decision Provenance
Every fraud decision backed by supply-chain-grade evidence: input signals, model version, policy version, override history. Built for regulators and disputes.
Remediation Language
Deterministic, auditable response actions -- quarantine, step-up, hold, callback -- not open-ended AI decisions. Safe automation your compliance team can trust.
Pricing

Enterprise Intelligence. Startup Pricing.

NICE Actimize starts at $500K/year. SEON starts at $7,200/year. Fraudeny starts at $0.

Free
$0/mo
Get started. No credit card.
500 requests/day
Basic scoring
API docs + Swagger
MOST POPULAR
Starter
$99/mo
Full scoring + explainability.
10,000 requests/day
Full scoring + SHAP
Webhook alerts
Email support
Pro
$499/mo
Scale with confidence.
100,000 requests/day
Everything in Starter
Batch scoring API
Slack + priority support
Enterprise
Call
Tailored to your scale.
Unlimited requests
Custom ML models
Dedicated engineer
On-prem option + SLA
Amir Shachar - Founder of Riskernel
About the Founder

Built by Someone Who Shipped Fraud Systems

"I spent years building fraud models for large institutions. Riskernel is the product I wanted when speed, explainability, and ops usability all mattered at once."

Former AI Research Team Leader at NICE Actimize (4 years)

Former Chief Data & AI Scientist at Skyhawk Security

Background across fraud detection, cybersecurity, and ML infrastructure

Author of the Semi-discrete Calculus and Algogens AI theories

Featured in TechTarget, Dark Reading, CSO, Yahoo Finance

View Patent List
Work With Amir

Make the expert surface explicit

If someone finds Riskernel because of the writing, they should immediately understand the next move: book Amir for a podcast or conference, use him for a paid vendor-evaluation or shadow-test decision, or follow the weekly briefing and quarterly report.

Speaking + Podcasts

For podcasts, conferences, panels, and internal briefings that need an operator voice on fraud infrastructure, vendor evaluation, explainability for ops, payout risk, and real production tradeoffs.

View topics and booking details

Expert Advisory

For paid expert sessions on vendor evaluation, shadow-test design, payout-control reviews, and the operational decisions that should be made before a team commits.

See advisory use cases

Weekly Briefing

A weekly note for serious fintech operators. No generic founder commentary, just concrete notes on fraud infrastructure, rare-event modeling, payout risk, latency, and explainability.

Join the briefing

Quarterly Fraud Infrastructure Report

A sharper public artifact for podcast hosts, conference organizers, and buyers who want one place to understand the current operator point of view.

Read the current report

Every Minute Without Fraud Detection Costs You Money

The average fraud attack costs $4.24 per dollar lost. Start protecting your transactions in the next 5 minutes.

Join the waitlist for early API access. No credit card required.