FTD Risk Intelligence for Online Payments

Your MID is your business.
We keep it alive.

A real-time AI risk indicator for specialized acquirers and PSPs. Every first-time deposit is scored before it touches your MID — so you catch the risky traffic that burns MIDs, and act on it before it settles.

Get Free Risk Exposure Analysis → See the API No contracts · Live in days
0
Data points the AI engine analyzes on every FTD decision
0
Average decision latency, per FTD, via one API call
Millions
Of the FTDs our models are trained on — and counting

FTDs are where MIDs go to die

First-time deposits are where fraudsters test stolen cards. No history, no trust signals — and every fraudulent FTD that settles counts against your scheme ratios twice.

🔍

No History

FTDs carry zero trust signals. Fraudsters use them to test stolen cards against your MIDs before moving real money.

Highest-risk txn class
⚠️

Counted Twice

A fraudulent FTD hits your ratio twice: once as a TC40 fraud report, and again as a dispute when the chargeback lands.

Double exposure

Portfolio Risk

One bad merchant or traffic source can drag your whole portfolio over scheme thresholds and put every MID at risk.

Contagion

A burned MID isn’t a fine — it’s a lost banking relationship, frozen reserves, and merchants you can no longer board.

Card-network compliance, built in

Visa and Mastercard hold acquirers to strict integrity standards. Staying compliant means keeping fraudulent and non-compliant traffic off your MIDs in the first place.

The Standard

Programs like Visa’s VIRP and Mastercard’s BRAM require acquirers to keep illegal, prohibited, and fraudulent activity off their MIDs — transaction laundering, prohibited goods, miscoded MCCs. Violations can mean significant fines, MATCH listing, and MID termination.

🛡️ How MIDShield helps

We help you stay genuinely clean. The AI engine flags fraudulent and non-compliant FTDs in real time so you can keep them off your MIDs entirely — a portfolio that meets card-network standards by design. Real-time compliance, not after-the-fact firefighting.

One API call. One AI decision. Clean ratios.

Built exclusively for specialized acquirers and PSPs — AI models trained on millions of FTDs, not generic e-commerce traffic.

01

Route

Every FTD is sent to the MIDShield API before authorization.

02

AI Analysis

ML models extract 2,000+ features: device graph, BIN intelligence, velocity, behavior.

03

Decide

APPROVE / DECLINE / CHALLENGE (3DS) — in under 100ms.

04

Learn

Every decision retrains the models. The network gets smarter with every transaction.

midshield · decision-api POST /v1/decisions
# Authenticated request
POST https://api.midshield.ai/v1/decisions
Authorization: Bearer sk_live_••••4f2a
Idempotency-Key: "d9f1c0a7-..."
{
  "transaction_type": "FTD",
  "merchant_id": "mid_88213",
  "amount": 250.00, "currency": "EUR",
  "card": { "bin": "453201", "last4": "7781",
    "scheme": "VISA", "funding": "prepaid" },
  "customer": { "email_age_days": 3,
    "country": "DE", "is_returning": false },
  "device": { "fingerprint": "fp_8c2e…",
    "ip": "185.220.x.x", "proxy": true }
}
Response · 74 ms · HTTP 200
{
  "decision": "DECLINE",
  "recommended_action": "do_not_settle",
  "risk_score": 94, "confidence": 0.97,
  "signals": [
    { "code": "proxy_detected", "weight": 0.34 },
    { "code": "bin_velocity_spike", "weight": 0.29 },
    { "code": "device_linked_to_cb", "weight": 0.21 }
  ],
  "decision_id": "dec_01HZ…",
  "tokens_remaining": 84211
}
~74ms p95 latency REST + Webhooks 99.95% uptime SLA PCI-DSS scoped

Big data in. One decision out.

Every FTD decision is the output of models that learn from the entire network — not static rules that fraudsters reverse-engineer in a week.

🧠

ML models, not rule lists

Gradient-boosted and graph-based models trained on millions of FTDs, retrained continuously as fraud patterns shift — no manual rule maintenance.

📊

2,000+ features per decision

Device graph, BIN intelligence, IP reputation, velocity vectors and behavioral telemetry — extracted and scored in real time, under 100ms.

👁️

Behavioral analytics

Session telemetry and biometric patterns score user intent — separating genuine depositors from scripted card-testing.

🛰️

Proxy / VPN / TOR

Network anonymization and true-location detection unmask actors hiding behind residential proxies and VPNs.

🌐

Big-data network effect

Every decision feeds the data flywheel. A fraudster flagged at one PSP is recognized across the entire MIDShield network instantly.

Real-time, every time

One synchronous API call returns a decision in ~74ms — fast enough to sit inline before authorization without slowing checkout.

Generic fraud tools protect merchants. We protect acquirers.

We don’t compete with merchant-side tools — we sit in front of the MID, where the acquirer’s risk lives.

Them (Forter, Ravelin…)
MIDShield
Built for
Merchants & e-commerce checkout
Specialized acquirers & PSPs only
Optimizes for
Conversion rate
MID survival & scheme ratios
Specialization
Generic transaction scoring
FTD-trained models
Commercials
Enterprise contracts, long cycles
Prepaid tokens — live in days, no lock-in

Pay per decision. Tokens never expire.

Buy tokens, plug in, and act on FTD risk before it settles. 1 token = 1 FTD decision.

Starter

10K tokens
$0.50 / decision
  • Full decision API
  • Dashboard & monitoring
  • Email support
Start Pilot

Scale

1M tokens
$0.30 / decision
  • Everything in Growth
  • Per-MID model tuning
  • Dedicated Slack/Telegram
  • Quarterly risk reviews
Talk to Sales

Enterprise

Custom volume
Custom pricing
  • Dedicated models
  • SLA & uptime guarantees
  • On-prem / VPC options
  • Named risk analyst
Contact Us

A token costs cents. A burned MID costs everything.

How exposed is your portfolio?

Free risk exposure analysis: send us 60 days of anonymized FTD data — we’ll show you exactly which transactions would have been declined, and where your ratio would be with MIDShield in front of it.

Message Us on Telegram →

Telegram: @midshield  ·  sales@midshield.ai  ·  No contracts. No lock-in.

Gulliver Creative Quarter, 1 Sportyvna Square, Kyiv 01023, Ukraine