Catch the ring before it onboards.
Every application is scored against a per-tenant identity graph. Duplicate identities, shared devices and fraud rings surface as explainable signals — with thresholds your team controls.
The Identity Graph
One graph, every applicant, every link.
Each application contributes nodes to a per-tenant graph — email, phone, device fingerprint, IP and address, plus hard identifiers like passport, national ID, PAN and business registration. Attribute values are hashed, so links are found without storing raw identifiers. When applications share a node, the graph connects them.
Three ways fraud gives itself away.
Each signal runs in order and contributes to the score. Every one is explainable — you see exactly which fired and why.
Duplicate Identity
Flags when a hard identifier — passport, national ID, PAN, SSN, tax ID or business registration — already appears on other applications.
Shared Device
Flags when a device fingerprint is reused across multiple applicants, a classic signal of coordinated or synthetic onboarding.
Identity Ring
Walks the graph up to three hops from the applicant to count linked applications, surfacing fraud rings that share attributes indirectly.
A single score from 0 to 1000.
Signals compose into one composite score, mapped to a level and a recommended decision. Approve clean applicants automatically; route the rest to review.
Low
0–249
Approve
Medium
250–499
Review
High
500–749
Manual review
Critical
750+
Reject
720
/ 1000 · High
Signals fired
- Duplicate identity+200
- Shared device+180
- Identity ring (3 linked)+150
Tuned to your risk appetite.
No hard-coded thresholds. Map your onboarding data to fraud attributes and tune how aggressively each signal fires — all per tenant.
Fraud Data Mapping
Point each fraud attribute at a built-in field, a request signal or a custom onboarding field, and mark which are hard IDs.
Per-Signal Sensitivity
Set flag-after-N-applicants, points-per-hit and a maximum cap for each signal independently.
Enable What You Need
Turn each check on or off. Defaults flag on any reuse, so you start strict and relax deliberately.
Where it runs
Onboarding and payments, covered.
The identity graph scores applicants the moment they register. On the payment side, the deterministic fraud scorecard and transaction-ring detection run inside the monitoring pipeline — so the same fraud lens covers onboarding and every transaction.
- Scored on registration, via API or SDK
- Explainable signals on every result
- Feeds the review queue and webhooks
Payment-side fraud
See it score your own applications.
Bring a sample and we'll show the graph linking what should never have been linked.