Documentation

RiskGuard Developer Docs

Everything you need to integrate real-time fraud detection into your fintech application.

Quick Start

quickstart.sh

# Install the SDK
npm install @riskguard/node

# Or use curl directly
curl -X POST https://api.riskguard.wejoona.com/v1/evaluate \
  -H "Authorization: Bearer rg_live_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "transaction_id": "txn_001",
    "amount": 150000,
    "currency": "XOF",
    "sender": "usr_abc",
    "receiver": "usr_xyz",
    "channel": "mobile_money"
  }'

Start here

Getting Started

Create an account, get your API key, and evaluate your first transaction in under 5 minutes.

Reference

API Reference

Full REST API documentation. Endpoints, request/response schemas, error codes, and authentication.

Guide

Rule Engine

Write velocity rules in YAML. Thresholds, time windows, geo-fencing, and custom conditions. Hot-reload without deploys.

Guide

Graph Analysis

How RiskGuard uses Neo4j to detect fraud rings, mule chains, and collusion patterns. Includes Cypher query examples.

Core Concepts

Evaluation

A single risk assessment of a transaction. You POST transaction data, RiskGuard returns a score (0.0-1.0) and an action (allow/review/block). Each evaluation runs through all enabled layers.

Rule

A YAML-defined condition that checks transaction attributes. Rules can reference velocity windows (e.g., "count of transactions from this phone in the last hour"), amounts, metadata, and time-of-day.

Graph Query

A Neo4j Cypher pattern that detects structural fraud. Example: "find all users who share a device with a flagged user and have transacted with the same merchant in the last 7 days."

Watchlist Hit

A match against sanctions or PEP lists. Fuzzy matching handles name transliteration. Each hit includes a confidence score and the matched list entry.

Decision

The final output: allow, review, or block. Includes the score, which rules fired, any watchlist hits, and graph patterns detected. Stored in PostgreSQL for audit.