Skip to content
IPRF

Instant payments. Higher trust.

Faster paymentssafely, by design.

IPRF is an open-source framework for fraud prevention and operational resilience in instant payment systems.

Modular, transparent, and built for low latency, IPRF helps financial institutions evaluate risk in real time — so legitimate payments flow, and fraud is stopped before it happens.

Explore the FrameworkView on GitHub
  • Open Source
  • Modular & Extensible
  • Reference Implementation
  • Real-time decisions
  • Low latency
  • Higher trust
  • Greater resilience

Scroll to follow a transaction

Follow a transactionthrough multiple layersof protection.

From identity to post-settlement analysis, IPRF combines deterministic rules, pre-computed risk intelligence and asynchronous enrichment across multiple layers — without slowing down legitimate payments.

  1. Transaction start

    A payment is initiated and received by IPRF in milliseconds.

    {
  "amount": 125.00,
  "currency": "USD",
  "from": "user_123",
  "to": "user_987"
}

  2. Layer 1

    Identity & Account Posture

    We verify identity, account history, device, channel and account age to assess the trust level of the payer.

    in-path< 1 ms
    • Identity verification
    • Account reputation
    • Device and channel analysis
    • Historical behavior

  3. Layer 2

    Real-Time Behavioral Scoring

    We analyze amount, counterparty, timing, channel and velocity using deterministic rules with strict latency budgets.

    in-path< 5 ms
    • Amount and pattern analysis
    • Velocity checks
    • Behavioral risk scoring
    • Real-time decision rules

  4. Layer 3

    Counterparty & Network Signals

    We evaluate counterparty risk and network intelligence using pre-computed risk state — never live database lookups.

    in-path< 5 ms
    • Counterparty risk score
    • Network relationship analysis
    • Known fraud patterns
    • Pre-computed risk state

  5. Layer 4

    External Enrichment

    We asynchronously enrich the transaction with external intelligence, sanctions data and contextual signals.

    asyncBackground
    • External data sources
    • Sanctions and watchlists
    • Fraud intelligence providers
    • Context enrichment

  6. Layer 5

    Post-Settlement Analysis

    We analyze completed transactions to detect new patterns and feed future decisions.

    asyncContinuous
    • Pattern detection
    • Typology analysis
    • Feedback to risk models
    • Continuous improvement

  7. Decision

    The transaction is allowed, sent for review, or declined — with a clear, explainable reason.

    • Allow
    • Review
    • Decline

    Fast. Explainable. Auditable.

Built for performance. Trusted for what matters.

  • < 10 ms

    In-path latency budget (p95)

    TARGET

  • 5 layers

    Three in-path, two asynchronous

    FACT

  • Deterministic

    Rules and thresholds, no black box

    FACT

  • 100%

    Decisions written to an audit trail

    OPEN CORE

Target is a design budget the framework is built to. Fact is a property of the source you can verify by cloning it. No figure on this page is a benchmark result — the suite has not been run.

Open source. Real impact.

Integrate. Adapt. Scale.

Get the code, explore the documentation and see how IPRF can fit into your payment infrastructure.

Explore on GitHubRead the Documentation
RequestPOST/api/v1/transactions/evaluate
{
  "transactionId": "txn_123",
  "amount": 125.00,
  "currency": "USD",
  "from": "user_123",
  "to": "user_987"
}
Response200 OK
{
  "decision": "ALLOW",
  "riskScore": 0.12,
  "latencyMs": 0.41,
  "riskFactors": [],
  "frameworkVersion": "0.1.0-SNAPSHOT"
}

Synthetic / demo data — illustrative payload, not a recorded transaction

Safer payments. Stronger systems.

IPRF is open source and free to adopt, adapt and audit. If you are assessing fraud controls on an instant-payment rail, start with the methodology.

Get in TouchExplore GitHub