📘 Public beta · Endpoints are stable; OpenAPI specs and SDKs ship monthly. See changelog →
Products
Anti-Fraud Platform
Quickstart

Anti-Fraud Platform · Quickstart

End-to-end: register a device session (simulated, no SDK), evaluate a transaction, inspect the resulting alert.

Time: ~10 minutes.

Prerequisites

Sandbox API key with scopes: evaluate:write. (For dashboard work, you'd need broader scopes; evaluate:write covers the programmable path.)

1. (Optional) Register a device session

If your evaluate call originates from a mobile app, the SDK collects device signals and posts them here, getting back a sessionToken you forward to evaluate.

For the quickstart we'll simulate the SDK with a raw POST:

POST/api/device-session
Auth · API keyScope · evaluate:write
curl -X POST https://sandbox.quantumelixir.tech/anti-fraud/api/device-session \
  -H "Authorization: Bearer $QE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "device": {
      "platform": "iOS",
      "osVersion": "17.4",
      "model": "iPhone15,4",
      "locale": "id_ID",
      "timezone": "Asia/Jakarta"
    },
    "riskSignals": {
      "jailbroken": false,
      "rooted": false,
      "emulator": false,
      "vpn": false
    }
  }'

Response:

{
  "data": {
    "sessionToken": "ds_01HXY...",
    "expiresIn": 1800,
    "attestationVerdict": null
  },
  "ok": true
}

You have 30 minutes to use this token in an evaluate call.

2. Evaluate a transaction

POST/api/evaluate
Auth · API keyScope · evaluate:writeRate limit · 100/min default · up to 20k/min
curl -X POST https://sandbox.quantumelixir.tech/anti-fraud/api/evaluate \
  -H "Authorization: Bearer $QE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "customerId": "cus_demo_12345",
    "lane": "transaction",
    "sessionToken": "ds_01HXY...",
    "externalId": "tx-2026-05-24-001",
    "amount": 250000,
    "currency": "IDR",
    "data": {
      "channel": "qris",
      "beneficiaryName": "Warung Sate Pak Budi",
      "merchantCategory": "5812"
    }
  }'

Response:

{
  "data": {
    "decision": "allow",
    "fraudScore": 12,
    "subScores": { "rules": 0, "velocity": 5, "ml": 7 },
    "appliedRules": [],
    "sessionId": "alt_01HXY..."
  },
  "ok": true
}

sessionId is the alert ID if an alert was created. For a benign transaction like this, the alert exists but decision: allow means no action is needed.

3. Force a block

Try an amount that hits the default high-amount rule + a sanctioned beneficiary in sandbox:

curl -X POST .../api/evaluate \
  -H "Authorization: Bearer $QE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "customerId": "cus_demo_12345",
    "lane": "transaction",
    "externalId": "tx-2026-05-24-002",
    "amount": 80000000,
    "currency": "IDR",
    "data": {
      "channel": "transfer",
      "beneficiaryName": "John FraudList",
      "beneficiaryAccount": "1234567890"
    }
  }'
{
  "data": {
    "decision": "block",
    "fraudScore": 92,
    "subScores": { "rules": 70, "velocity": 12, "ml": 10 },
    "appliedRules": [
      { "code": "BLK-001", "name": "Blocked beneficiary list", "action": "block", "severity": "critical", "score": 50 },
      { "code": "VL-003", "name": "High amount unverified counterparty", "action": "review",  "severity": "high",    "score": 20 }
    ],
    "sessionId": "alt_01HXZ..."
  }
}

decision: block → reject the transaction in your transaction processor.

4. List recent alerts

curl "https://sandbox.quantumelixir.tech/anti-fraud/api/alerts?status=open&lane=transaction&limit=20" \
  -H "Authorization: Bearer $QE_API_KEY"

Returns alerts your evaluations created. Use this for backoffice triage UIs or to subscribe to a webhook.

Production wiring

In production, two pieces shift:

  1. SDK does step 1. Your mobile app embeds @quantum-elixir/device-sdk-rn and calls device-session from the device. You never assemble the payload server-side.
  2. Evaluate becomes inline-blocking. Your transaction processor calls /api/evaluate before committing. p95 < 100ms means it can sit on the critical path.

Next steps

OverviewRisk evaluation