Abuse Triage &
Enforcement
From “something's wrong” to “it's handled” without leaving the dashboard.
Book a demoA ranked queue of every account, sorted by risk, filterable by tier, label, and subscription. Open one to see its score history, 30-day token/cost rollup, model mix, and the exact signals that fired. Confirm a verdict and block the account in one click. Every label you submit feeds the next nightly retrain, so the model sharpens on your abuse.
- Cost & consumption
- Token-velocity bursts, robotic timing, instant-first-call after signup, cheap-model concentration, tiny-prompt→huge-completion ratios, 24/7 activity.
- Network
- Datacenter-IP origin (AWS/GCP/Azure/OVH/Hetzner/…), IP rotation, and sybil clusters (many accounts, one IP).
- Behavior
- Inter-request gap distributions, session fragmentation, day-of-week diversity (always-on automation vs. human diurnal rhythm).
Capabilities
Risk-ranked account queue + full per-account forensic detail
One-click block / unblock; blocked accounts are denied at
/v1/scoreGround-truth labeling (
fraud / legitimate) drives per-customer retrainingLive detection log: fingerprint, origin, IP, reasons, raw signals per event