Skip to content
LR TraceDevice Fingerprinting

Stop repeat fraud and abuse without adding login friction.

Know when the same physical device returns—even across browsers, cleared cookies, or private windows—so you can block multi-accounting, catch stolen credentials faster, and step up MFA only when risk is real. Stable IDs in milliseconds, not another hoop for good customers.

90+

Signals per identification

<40ms

p95 end-to-end latency

4

Major browsers, one device ID

197+

Tamper & API checks

Capabilities

Everything you need

DRAWNAPART GPU Timing

Inspired by DRAWNAPART research, our GPU timing analysis uses 10K+ vertex workloads to extract hardware-specific rendering signatures that distinguish individual GPUs.

Attention-Based ML Matching

Neural attention model weighs 37+ signal dimensions to produce a confidence-scored device match, continuously improving through our real ML training pipeline.

Canvas Fingerprinting

Sub-pixel rendering differences in HTML5 Canvas create unique signatures based on graphics hardware, OS, and font rendering engines.

AudioContext Analysis

Audio processing characteristics of the device's sound hardware produce consistent, cross-browser identifiers.

Cross-Browser Consistency

The same device ID is generated whether the user visits from Chrome, Firefox, Safari, or Brave. Our core differentiator.

Incognito Persistence

Hardware-level signals don't change between normal and private browsing sessions, making our fingerprints persistent across modes.

Signal Drift Detection

Automated monitoring detects when GPU, WebGL, or audio signals drift over time due to browser updates, ensuring accuracy as the landscape evolves.

Lies & Tampering Detection

Scans 197+ browser APIs for prototype tampering, spoofed properties, and fingerprint resistance — detecting manipulation attempts in real time.

Headless & Automation Detection

Identifies headless browsers, Puppeteer, Playwright, and automation frameworks through engine-specific signal analysis.

Behavioral Biometrics

Captures mouse dynamics, keystroke timing, scroll patterns, and touch gestures to strengthen identification and detect bot-like behavior.

Sub-40ms Latency

Signal collection, hashing, and server-side matching complete in under 40 milliseconds. Zero impact on page load times.

How It Works

Three steps to get started

01

Install the SDK

Ship identification in one dependency. Full TypeScript types so your team wires sessions and user IDs with compile-time safety.

02

Call identify()

On each visit, the SDK collects GPU, canvas, audio, WebGL, tamper checks, and behavioral signals—no prompts, no extra clicks for legitimate users.

03

Act on the verdict

Use device ID, confidence, and risk in your rules engine: allow, challenge, or block—typically in under 40ms end-to-end.

Live Playground

Your device, identified right now

The JSON on the right is the actual fingerprint just collected from your browser — no fixtures, no mocks. Switch languages to see the call that produces it.

import { LightningResearch } from "@lightningresearch/sdk";

 

const { sessionToken } = await fetch("/api/collector/session", {

method: "POST",

}).then((r) => r.json());

 

const client = new LightningResearch({

sessionToken,

endpoint: "https://api.lrdefender.lightningresearch.ai",

});

 

const result = await client.identify();

// → { deviceId, uniqueness, elapsed, signals, hashes }

200 OK · live from your browser
Collecting 30+ signals from your browser…
Collected in ·0 signals·Full API reference

Use Cases

Built for real-world security

Payment & signup fraud

Link chargebacks and stolen cards to the same hardware even when fraudsters rotate emails, browsers, and incognito windows—so risk teams review fewer false leads.

Trust scoring at login

Treat unrecognized devices as higher risk and trigger MFA or device binding only when Trace says the session looks new or inconsistent—not on every visit.

Promo & trial abuse

Stop one person from farming referral credits, trials, or signup bonuses by detecting repeat devices behind fresh accounts.

Marketplace & seller integrity

Surface sock-puppet sellers and coordinated fake buyers by resolving multiple storefront logins to the same device footprint.

High-value account recovery

When a user resets a password or moves money, compare the requesting device to history to block takeover flows that look like new hardware.

Get started with LR Trace

Free tier includes 10,000 identifications per month. No credit card required.