Autonomous AI agents that analyze every transaction, bet and session on your platform. Detect fraud, money laundering and abuse before they cause damage.
Each agent specializes in a threat category. Together, they form a complete barrier against fraud, abuse and regulatory risk.
Monitors suspicious financial patterns — disproportionate withdrawals, smurfing, round-tripping and structuring.
Detects changes in player behavior — anomalous hours, device switching, bot patterns.
Identifies match fixing, coordinated betting and sharp money in sports markets.
Detects game exploits, bonus abuse, automation and suspicious grind patterns.
Ensures regulatory compliance — incomplete KYC, PEP matching, COAF limits and responsible gaming.
From 0.0 to 1.0, every transaction is automatically classified. You configure the thresholds and actions for each level.
Every event passes through multiple analysis layers. From simple rules to cross-player intelligence.
Operator-configurable thresholds. E.g.: "Withdrawal > 5x the 30-day average" triggers an immediate alert.
Two ways to connect your platform. Choose what works for you.
Direct connection, zero friction. Just authorize and all events flow automatically — no code, no deploy.
// Softswiss
> Authorize → Connect → DoneSame SDK as Commissions and Geolocs. Enable the anomalies module and start sending events with one line of code.
$ npm install @rakero/sdk
// Enable module
rakero.anomalies.track({
type: 'bet_placed',
playerId: 'player_123',
amount: 50000
})
Zero friction integration. Want your platform integrated? Get in touch
Each agent monitors a threat category (financial, behavioral, sports, casino, compliance). They analyze events in real time, compare with the player's historical profile and generate an anomaly score (0.0 to 1.0). Scores above the configured threshold trigger alerts or automatic actions.
The operator configures actions by risk level: alert the team, block withdrawal, limit stakes, force KYC verification, or pause the account. Everything is configurable per agent and per threshold.
Yes. For Softswiss, integration is native via Kafka (zero code). For other platforms, we offer the @rakero/sdk which is already used by Commissions and Geolocs — just enable the anomalies module.
The architecture is based on ClickHouse and Go, designed to process millions of events per month per operator. Each operator has an isolated database (multi-tenant), ensuring performance and compliance.
Each event receives an anomaly score from 0.0 to 1.0. Normal (0-0.3) is logged. Suspicious (0.3-0.6) shows in the dashboard. High risk (0.6-0.8) triggers an alert. Critical (0.8-1.0) can trigger automatic action.
Yes, that's the key advantage. Anomalies cross-references data from Commissions (fraudulent affiliates) and Geolocs (location contradicts session) for cross-product detection that no competitor offers.