Fingerprint for Stripe

The fraud signal Stripe Radar
never had.

Persistent device identity that survives card changes, new accounts, and VPN rotation. Built for payments teams that know Radar is only the beginning.

Book the walkthrough
99.5%
ID Accuracy
0.1%
False Positive Rate
4ms
API Response

Why Stripe teams are adding Fingerprint now.

Stripe processes $817B in payments annually. Radar's rule engine is excellent at catching known fraud patterns — but rule-based systems have a structural blind spot. When a fraudster resets their card, creates a new account, or rotates through a VPN pool, Radar sees a fresh identity. Fingerprint doesn't. Our device signal is derived from dozens of browser and hardware signals, producing a persistent visitor ID that remains stable for months or years regardless of cookies, IPs, or card numbers. That's the signal Radar never had — and the one that closes the gap on device-level payment fraud, account takeovers, and bot-driven credential stuffing at scale.

Four signals. One API call.

Persistent Device ID

A stable visitor identifier that survives cookie clears, incognito mode, VPN rotation, and card resets. 99.5% accurate across the open web.

Payment Fraud Detection

Device-level conviction scores passed directly into Stripe's charge metadata. Flag suspicious devices before Radar processes the transaction.

Bot Protection

Detect automated fraud attempts, credential stuffing, and card testing bots before they generate chargebacks or trigger false positives in Radar.

Account Takeover Shield

Surface logins where the device doesn't match the account's trusted history — catching ATO attempts that look clean from an IP and card standpoint.

"It acts as fraud detector and a trust enabler."
Andreas Zodhiates, Head of Payment Fraud, Booking.com

Booking.com integrated Fingerprint to close the device-level gap in their payments stack. The result: reduced chargebacks from repeat fraudsters and improved checkout conversion for trusted customers, because the device signal lets them reserve friction for sessions where fraud risk is real — not just unfamiliar IPs or new cards.

The math on prevention at Stripe's scale.

At $817B in annual GMV, each basis point of fraud represents roughly $82M. Fingerprint catches the fraud Radar can't reach — the device-persistent kind.
Fingerprint's false positive rate is 0.1%. Good customers don't get flagged. Revenue doesn't leak from friction-induced abandonment.
Cost as a percentage of prevented fraud is under one basis point. Ships in hours. No infrastructure changes. Works alongside Radar, Sigma, and existing fraud tooling.