Reference

Device intelligence glossary

Plain-English definitions for the fingerprinting, bot-detection and fraud-prevention terms you’ll meet across Prynt.

visitorId
A stable identifier for a browser or device, derived from many signals rather than a cookie, so clearing cookies or opening a private window does not reset it. It is scoped to your account and is the anchor for recognizing returning devices for fraud and abuse prevention.
Confidence score
A 0–100% value shipped with every identification, telling you how closely this visit matches the device’s known signals — so you can set different thresholds per action.
Smart Signal
A discrete risk indicator returned alongside identity: bot, incognito, VM, VPN, Tor, datacenter, proxy, residential-proxy, tampering, location spoofing and more.
Canvas fingerprint
A signal derived from how a device renders text and graphics to an HTML canvas — subtle differences in GPU, driver and OS produce a highly device-specific value. Prynt uses a perceptual hash so small driver changes do not break the match.
JA4 fingerprint
A fingerprint of the TLS client hello, captured at the network edge before any JavaScript runs. A browser User-Agent riding a non-browser TLS stack (curl, Go, Python) is a spoofed client.
Residential proxy
An anonymizing network that routes traffic through real consumer IP addresses to look like an ordinary home user — a favorite of fraud rings because it evades naive datacenter-IP blocking. Prynt detects it explicitly.
Behavioral biometrics
Analysis of how a person interacts — mouse movement, keystroke timing, scroll and motion — to distinguish humans from bots and to catch account takeover on a victim’s own device.
Sealed result
An identification result encrypted (AES-GCM) on the client and decrypted server-side, so the verdict cannot be read or tampered with in transit.
App Attest / Play Integrity
Platform attestation APIs (Apple App Attest, Google Play Integrity) that cryptographically prove a mobile request comes from a genuine, unmodified app on real hardware.
Reputation network
A shared feed of one-way, salted hashes of entities that accounts confirmed as fraud, plus public blocklists. A confirmed fraudster is flagged on the next site it tries; identities are never shared, and any account can stop contributing.
Reason code
A short machine-readable label (e.g. VPN, TOR, AUTOMATION_BEHAVIOR) explaining why a session got its score, so your rules can act on specifics.
Suspect score
A single normalized risk value (0–1) summarizing all signals, used to drive allow / challenge / block decisions.

Keep reading

Device fingerprinting explained · Bot detection · Account takeover · Prynt vs FingerprintJS · SDKs