Temporal links
A connection weakens as it becomes stale, with half-lives suited to the underlying evidence type.
Temporal identity graph
Connect accounts throughevidence that ages.
Discover multi-accounting, referral rings and fraud neighbors through account-to-entity edges that account for rarity, recurrence and time, not simply whether two values ever matched.
Live evidence
8 linked
Current analysis
Identity graph
Account84%
Device62%
Email91%
RiskAccountDeviceEmailEvidence
Evidence attached
7+
entity types connected
Decay
per evidence category
Bounded
graph exploration
Built for the full investigation
Useful to engineering. Clear to fraud operations.
The same evidence powers API decisions and investigator views, so policy changes stay connected to the behavior behind them.
Rarity weighting
A rare shared browser installation matters more than a common ASN or office subnet.
Fraud-neighbor context
Known labels propagate useful context into decisions without turning proximity into automatic guilt.
Investigator workflows
Explore bounded neighborhoods, compare two accounts, identify connected components and open cases.
The operating flow
Move from signal to action without losing the explanation.
Authenticate a session
Your server associates the anonymous browser session with a canonical account identifier.
Write explainable edges
Device, browser, IP, subnet, ASN and email relationships are stored with strength and evidence.
Recompute meaningful links
Rarity and decay turn raw shared entities into an account-link confidence investigators can inspect.
Start with one protected workflow
Find coordinated behavior hidden between individual events.
Give fraud teams a graph they can explain, query and use directly in risk decisions.