Leaked card detection
Card sightings are Luhn-validated and matched against your registered BIN ranges in a database-backed registry — so the alert is about your cards, with the noise of invalid numbers and other banks' ranges already removed.
PANs hashed, never stored
Card numbers are SHA-256 hashed at ingestion with only first-six and last-four retained for matching. The leak doesn't get a second life inside your monitoring tool.
Sighting lifecycle
Every leaked card moves through a state machine — sighted, confirmed, actioned, closed — so reissue workflows and fraud teams work from the same authoritative status.
Custom DLP policies
Define regex policies for what counts as yours — code signatures, document markers, account formats. Policies are statically checked for catastrophic patterns and capped at 0.5s runtime with auto-disable, so a bad regex can't take the pipeline down.
Breach SLA tracking
Operator-tunable thresholds decide when accumulating sightings constitute a notifiable breach. When the line is crossed, the clock and the evidence pack are already assembled.
Routed to who acts
Card batches route to fraud ops, DLP hits to security, breach evaluations to compliance — via email, Slack, Teams, PagerDuty, Opsgenie, or webhook.
Register what's yours
BIN ranges into the registry, DLP policies for your data formats, notification thresholds per severity. This is the matching baseline.
Sightings stream in
Card and data sightings arrive from the dark-web crawlers, paste monitoring, and underground channels — every sighting carries its source and capture evidence.
Validate and match
Luhn validation kills fake numbers; BIN matching scopes to your ranges; DLP policies score data hits. Only what's real and yours becomes an alert.
Track the lifecycle
Each confirmed item moves through its state machine with full audit history — who confirmed, what action, when closed.
Evaluate breach posture
The breach engine continuously evaluates accumulated findings against your SLA thresholds, so notification decisions are made on data, on time.
Under the hood