Priced against your volume
Screening volume, chain coverage and how often you call the API differ from desk to desk, so the rate is quoted against what you actually run rather than a published tier.
Wallet screening and on-chain risk intelligence
Risk intelligence for every wallet you touch.
Screen any address and get a verdict, along with the money trail behind it. Coverage runs to 40+ blockchains.
POST /v1/screen200 OK
{ "address": "0x7f36…be1b", "chain": "ethereum" }
{
"verdict": "FLAGGED",
"risk_score": 87,
"confidence": 0.91,
"evidence": {
"connected_to": "0x098b…f21c",
"list": "SDN",
"path_len": "multiple",
"path": "attached to the result"
}
}Indirect exposure to an SDN wallet. Example response, not from a live account.
BitcoinEthereumTronSolanaBNB ChainPolygonAvalancheXRPLitecoinDogecoinBitcoin CashMoneroZcashDashPolkadotCosmosStellarAlgorandEthereum ClassicArbitrumBaseOptimism
Bitcoin, Ethereum and Tron among 40+ blockchains.
$1.28T
of value followed through the transfer graph
27,697,319
value transfers indexed and attributed
5,416,087
addresses held in the graph
8,742,269
directed counterparty relationships
Built for the examiner
Evidence your examiner can follow.
We line up with the obligations you are held to.
FATFGLOBAL
Recommendation 15 screening expectations and the published red-flag typologies.
OFACUNITED STATES
Sanctions screening against the SDN list, snapshot-versioned so any decision can be replayed.
VARADUBAI
VASP wallet screening and the record-keeping obligations that follow it.
MiCAEUROPEAN UNION
Audit trails a CASP can hand to a supervisor without rebuilding them.
Pricing
Pricing fits the way you screen.
No price is printed here, because what you pay should follow what you run. We quote against your volume, and we carry the cost of getting you live.
Credits for the first integration
Your account opens with credits so wiring up the API, tuning thresholds and replaying old cases do not draw down a budget before anything reaches production.
What comes back
Every screen comes back as a decision.
Each result arrives with the reasoning attached, so an analyst can act on it without going away to build the case first.
Nothing connects this address to a sanctioned or illicit source.
The trail reaches a custodian, where one wallet holds thousands of customers and the funds stop being separable. The boundary is named in the result.
Exposure to a sanctioned or illicit source, direct or through intermediaries. The connecting path comes with it.
Analysis
We follow the money across multiple hops.
Value is tracked forward through the transfer graph and resolved to the entities behind it, so a result rests on the transactions themselves.
COVERAGE
Complete histories
Full transaction history is pulled around every designated address.
TRAVERSAL
Multiple hops
Indirect exposure is traced forward through the transfer graph and attributed back to its source.
ACCOUNTING
FIFO and haircut
Both run in parallel and are reconciled, so an exposure share can be checked against a second method.
ATTRIBUTION
Contract level
Thousands of counterfeit contracts impersonate the major stablecoins. Matching on a ticker symbol waves all of them through.
Typologies
Launderers reuse the same four shapes.
These structures turn up again and again in designated wallets, and each one leaves a signature in the graph that survives a change of address.
SEQUENTIAL RELAY
Peel chain
Value relays one-to-one through fresh addresses, each forwarding almost everything it takes in.
HIGH-DEGREE DISPERSION
Fan-out
A wallet splits funds across thousands of counterparties at once, betting nobody follows every branch.
CYCLIC RETURN
Round-trip
Value returns to the wallet it started from after passing through intermediaries under the same control.
CONVERGENT RECEIPT
Shared intermediary
A single address taking receipt from multiple unrelated sanctioned entities, with no exchange label attached to any of them.
Platform
Behavioural analytics, pointed at the chain.
The modelling stack behind zScreen is the one we already run on behavioural data at scale. Pointing it at the transfer graph changed the inputs, not the machinery.
Positive-unlabelled learning
Confirmed designations are the only hard labels that exist on a chain. The ensemble fits them against repeated random draws of everything else and votes on the result, so an address is never assumed clean for the sake of the arithmetic.
Entity-grouped validation
Addresses under one actor never straddle a train and test split, so the model cannot flatter itself by recognising a cluster it already memorised.
Calibrated confidence
Every verdict carries a calibrated interval derived from how complete the underlying history was, so a thin record never arrives looking like a full one.
Technique index
- MODELLING
- PU bagging, 200 modelsout-of-bag aggregationbeta calibrationVenn-Abers intervalsmonotonic constraints
- VALIDATION
- StratifiedGroupKFold 5x3TreeSHAP attribution53 features, 7 families
- GRAPH
- graph-native traversaltemporal orderingcluster co-attributionforward taint, time-ordered
- ACCOUNTING
- FIFO accountinghaircut accountingcustody-boundary cut nodes
- MATCHING
- homoglyph-resistant matchingcontract-level attribution
Inside one alert
Every alert explains itself in plain language.
The result says what it saw and which way each thing pushed the score, in wording an analyst and an auditor read the same way.
| What we saw | Effect | How sure |
|---|---|---|
| Money moved straight back out within a day | Raises risk | Strong |
| Funds passed along in a chain of one-to-one transfers | Raises risk | Strong |
| Most of the outgoing value went to an exchange | Raises risk | Moderate |
| The wallet has very little history behind it | Lowers risk | Weak |
An illustrative alert, in the wording an analyst sees on screen.
Bring a wallet you already have questions about.
Ready to get on a call?
Book a demo