Models · Fraud / AML

Fraud Graph

Graph-neural-network fraud detection for card-not-present and ACH transactions, with sub-50 ms inference at the edge.

  • Risk: low
  • GA
  • US
  • UK
  • EU
  • UAE
Recall @ 0.1% FPR
0.79
Median latency
32 ms
Graph nodes
1.2 B

What it does

Scores a transaction in real time using a graph constructed from device fingerprints, IP routes, merchant ids, and historical chargeback patterns. The scoring runs on Cloudflare Workers + Vectorize, so the round-trip stays under 50 ms even from regions with no major cloud presence.

Pricing

Available from Starter. Enterprise customers can run a dedicated graph in a private region.

Audit & jurisdiction

  • US — reviewed and routed to local counsel.
  • UK — reviewed and routed to local counsel.
  • EU — reviewed and routed to local counsel.
  • UAE — reviewed and routed to local counsel.

Every run of this model is appended to a hash-chained audit log. Anyone with a run id can fetch a Merkle inclusion proof — see the Compliance & Trust page.

Reported metrics are self-attested by the model author and verifiable against the audit log.