Prevent MEDs in your crypto operation
MEDs are chargeback claims on Pix, Brazil’s instant payment rail. Hetuva prevents them before the claim exists: real-time scoring and network intelligence across the ecosystem.
Request an analysis of your dataA MED is invisible until it’s too late
By the time the claim arrives, the fraud already happened — and every MED counts against the 1% line that sustains your banking relationship. Operations relying on isolated rules discover the pattern only after the third occurrence.
1% is the line. Counted by occurrences, not by value.
Bring every relevant signal into one risk context
Transactions, accounts, devices, sessions, wallets, disputes and external intelligence arrive as events: one continuous stream, correlatable from the first byte.
Find relationships isolated rules cannot see
The evidence graph connects accounts, devices and wallets with confidence-weighted edges — and highlights fraud communities while they form.
- 3 accounts
- 1 shared device
- funds dispersed in 4 min
Every risk result comes with evidence
No black box: every decision carries the reasons behind it — ready for audit, review and defense with your bank.
{
"risk_band": "high",
"confidence": 0.91,
"reasons": [
"linked_flagged_identity",
"rapid_fund_dispersion",
"device_account_reuse"
]
}How it works
- 1
Events
Your operation emits events in real time or in batch.
- 2
Evidence graph
Each event becomes correlatable nodes and edges.
- 3
Risk result
A score with explainable reasons, per transaction or entity.
- 4
Action
Webhooks and the API return the decision to your systems.
Capabilities
Identity correlation
The same person behind many accounts, seen as one.
Fund-flow tracing
Where the money went — even when it disperses fast.
Temporal patterns
Sequences and windows static rules cannot capture.
Explainable decisioning
Every score ships with the reasons that produced it.
Built for Pix-native operations
On/off-ramps
Protect fund entry and exit without blocking legitimate customers.
Crypto payments
Per-transaction risk with network context, not isolated rules.
Wallet platforms
Identify linked accounts before they become a MED cluster.
Real infrastructure
- Real-time assessment API
- Batch historical import
- Streaming ingestion
- Tenant-isolated data
- Webhooks and event delivery
Backtest results over real, anonymized Pix operations.
Designed for
- Shorter fraud investigations
- Consistent risk reasoning across analysts
- Earlier identification of linked entities
- Less dependence on isolated rules
See what we would have caught in your operation
We replay your history through our scoring — no integration, no commitment — and show what would have been prevented.