Introduction
Introduction
Learn the MaiGuard fundamentals, then move from quick start to production-grade authentication and integration patterns.
Tip
What is MaiGuard?
MaiGuard is a real-time trust layer for digital financial interactions. It sits between your transaction events and your business decisions — evaluating risk signals and returning a clear verdict so your platform can act immediately.
It evaluates signals across six dimensions:
- →Transaction behavior — amounts, velocity, and patterns
- →Device and network context — IP reputation, device fingerprinting
- →User and authentication signals — login patterns, session anomalies
- →Counterparty and beneficiary risk
- →Geo-location consistency
- →Historical account behavior
For each event it returns a riskScore, a decision (ALLOW / REVIEW / BLOCK), a set of explanatory flags, and optional downstream actions — all in under 200 ms.
Note
How it works
- 1Send
POST /v1/transactions/scorewith your transaction payload. - 2MaiGuard evaluates behavioral, network, and identity signals against your configured rule set.
- 3Receive
decision,riskScore, andflagssynchronously in under 200 ms. - 4Optionally submit feedback after the transaction resolves to continuously improve rule precision.
Pick your integration
Client-side
JavaScript SDK
Integrate behavioral biometrics and device intelligence directly into your web app. Collect signals silently, pass the session ID to the scoring API.
- →2-line quick start
- →Behavioral biometrics (keystroke, mouse, scroll)
- →Device fingerprinting & bot detection
- →React, Vue, and vanilla JS examples
Server-side
API Reference
Complete reference for all MaiGuard endpoints — REST, webhooks, database sync, SQS, and file imports. Everything you need to wire up fraud detection at any scale.
- →Transaction scoring & batch scoring
- →Webhooks: outbound alerts & ingestion
- →Database sync, SQS, and file import
- →Rate limits, error codes & retries
Implementation guide
Customer Onboarding Guide
Onboard customers, manage tenant lists, and align identity across onboarding and scoring calls.
- →PUT onboard, GET customers, tenant list entries
- →Identity alignment: same ID for onboarding & scoring
- →Errors, retries, and rate limit guidance
Implementation guide
Feedback Intelligence Guide
Close the loop — map scoring outcomes to feedback, improve rule precision, and monitor drift over time.
- →Single and batch feedback ingestion
- →Idempotency and safe retry strategy
- →Drift monitoring & phased rollout checklist
Integration channels
Six supported methods — from synchronous REST calls to scheduled database syncs and high-throughput queues.
REST API
1,000/min per tenant
JavaScript SDK
Behavioral biometrics
Webhooks
Event-driven
Database Sync
PostgreSQL · MySQL · MongoDB
AWS SQS
1,000 msg/s
File Import
CSV · JSON · JSONL
Quick start
Score your first transaction in three steps.
- 1Get your API key from the MaiGuard dashboard. Use a
pk_test_key during integration. - 2Call
POST https://api.maiguard.com/v1/transactions/scorefrom your backend with the transaction payload. - 3Read
decisionfrom the response —ALLOW,REVIEW, orBLOCK— and act accordingly.
Tip
pk_test_ to pk_live_ at go-live — no code changes required. Test-mode traffic is excluded from live reporting.Authentication
All API calls are authenticated via the Authorization header. Your tenant is resolved automatically from the key — never pass a tenantId in request bodies.
API key
Authorization: Bearer pk_live_<your-key>Server-to-server calls: scoring, imports, webhooks. Never expose in browsers or mobile apps.
JWT
Authorization: Bearer <access_token>Obtained via POST /v1/auth/login. Used for portal-style management routes (rules, customers, analytics).
Warning
pk_live_* keys in browser or mobile clients. Call the scoring API from your backend and pass only the decision to the client.Next steps
Was this page helpful?