For crypto casinos · Curaçao / Anjouan / CR

Stop bonus abuse.
Reclaim your margin.

Real-time anti-fraud built for iGaming. Send a player event, get a synchronous risk decision in under 100 ms — with the reasons you can defend in front of finance, ops, and your license.

Get sandbox access → See the API No KYC. No card fees. Crypto-native.
The cost of fraud at a crypto casino

Bonuses and withdrawals are leaking — silently.

Generic anti-fraud was built for card payments. We built ours for the actual fraud patterns that hit crypto-licensed iGaming operators every day.

01

Bonus abuse rings

Same device, same wallet, same card_hash across 8 accounts farming your welcome bonus.

02

Mixer-funded deposits

Wallets 4 hours old, funded straight from Tornado Cash or a sanctioned cluster. Untouchable on payout.

03

Account takeover

Password-reset → new device → withdraw, in 6 minutes. Your existing AML stack misses it entirely.

04

Affiliate self-dealing

Cohorts where 90% of "players" were created in the same hour, same /24, with the same fingerprint stack.

How it works

One POST. One synchronous decision. Three reasons.

No SDK lock-in. No 30-day batch jobs. Server-to-server, <100 ms p99, with explainable reasons your risk team can audit.

1

Send the event

signup · login · deposit · withdraw · bet · bonus_claim — six event types covering the full player lifecycle.

2

Get a decision

17+ rules + graph features + behavioral biometrics evaluated synchronously. Returns risk_score, decision, reasons.

3

Act on the reasons

Allow, hold for review, or block. Reasons are machine-readable so you can map them to your CRM, retention flow, or chargeback narrative.

curl
response
# withdraw event — server-side
curl -X POST https://api.antifraud.dev/v1/events \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operator_id": "op_3Yk...",
    "player_id": "plr_8721",
    "event_type": "withdraw",
    "timestamp": "2026-05-03T07:14:22Z",
    "amount_usd": 2400,
    "wallet_address": "0x9f2c...",
    "chain": "ethereum",
    "device_id": "d_a4f9...",
    "ip": "203.0.113.42"
  }'

# → 200 OK · ~38 ms
{
  "risk_score": 87,
  "decision": "block",
  "reasons": [
    "wallet_funded_by_mixer",
    "shared_device_with_4_accounts",
    "withdraw_without_play"
  ]
}
What we detect

17 detection modules. Every one explainable.

Rules, graph features, server & device fingerprinting, behavioral biometrics, wallet intelligence, and a privacy-preserving cross-operator network — all returning machine-readable reasons.

Graph fraud rings

Connected components on shared cards, IPs, devices, email domains. Communities surfaced with PageRank + Louvain.

Wallet intelligence

BTC, EVM, Tron, Solana: chain age, OFAC clusters, mixer routers, multi-account-by-wallet fan-out.

Device fingerprinting

JS SDK + server-side JA3/JA4 + header-order. Catches emulators, headless Chrome, and VM stacks.

Account takeover

Per-player baselines: new country, new device, geo-impossibility, recovery → withdraw chains.

Behavioral biometrics

Mouse, keystroke cadence, scroll, form-fill timing — GDPR-clean, opt-in, no values captured.

Proxy & IP intel

Residential proxy heuristics, Tor exit nodes, datacenter ranges. Pluggable upgrade to IPQS / Spur / Greip.

Affiliate fraud

Cohort analytics per affiliate: self-referral, bot traffic, cross-affiliate collusion, dedicated cohort model.

Sanctions & mixers

OFAC list, Tornado Cash, ChipMixer, sanctioned exchange clusters — refreshed daily, versioned audit trail.

Velocity rules

Withdraw without play, deposit-to-withdraw <5min, bonus-to-withdraw ratio, rapid-fire signups per /24.

Cross-operator network — the moat

Catch the same fraudster on day one — not your 90th day.

An opt-in, privacy-preserving blocklist shared between participating operators. Identifiers are hashed with rotating peppers via HKDF-SHA256 — no operator can rebuild another's data, and reporters are never disclosed in lookups.

  • HKDF-SHA256 with versioned peppers and per-identifier-type isolation
  • 3-bucket coarse hit-counts — never expose raw numbers
  • Write-only audit table; opt-in per-operator, revocable any time
  • 4-adversary threat model documented; right-to-erase honoured
net op_a op_b op_c op_d op_e op_f
Why us

Built for crypto casinos. Not retrofitted from card-payments.

We don't compete with KYC providers — we point you to Sumsub for that. We compete with generic anti-fraud and we win on the patterns specific to your business.

Antifraud
Sumsub
SHIELD
SEON
Crypto-native (wallet intel, OFAC, mixers)
✓ Native
— KYC only
Limited
Limited
Explainable graph fraud rings
✓ Built-in
Device-only
Add-on
Cross-operator network
✓ Privacy-preserving
iGaming-native rules (bonus abuse, withdraw-without-play)
✓ 17+
Generic
Generic
Synchronous decisions <100 ms
✓ p99
Async
"Fraud + Revenue" positioning
✓ Bonus ROI
17modules
From ingest to graph to ATO to affiliate cohort ML
<100ms p99
Synchronous decisions on the hot path
~750tests
API + Dashboard + SDK suites green
1–3days
Operator onboarding with Postman + SDK examples

Ship a sandbox integration this week.

Postman collection. OpenAPI 3.1. Code samples in Python, Node, PHP, Java. Onboarding without a sales call.