Fingerprint IDs

Neoprint generates 4 different fingerprint IDs, each optimized for a different use case.

Overview

const fp = await neoprint.get()

fp.id              // Full hash — maximum uniqueness
fp.stableId        // Survives browser updates
fp.weightedId      // Fewer collisions in corporate environments
fp.crossBrowserId  // Same across Chrome, Firefox, Safari, Edge

ID	Use case	Survives update	Cross-browser	Collision resistance
id	General identification	No	No	High
stableId	Long-term tracking	Yes	No	Medium
weightedId	Corporate/school	No	No	Very high
crossBrowserId	Cross-browser	Yes	Yes	Medium

id — Full Hash

Uses all available collectors. Maximum uniqueness but changes when any signal changes (browser update, new font installed, screen resolution change).

Best for: Short-term identification, session-level tracking.

stableId — Update-Resistant

Uses only the most stable collectors: math, webgl (vendor/renderer/params only), fonts, intl, gpu, cssFeatures.

speech is excluded because Safari returns 0 voices in private browsing, breaking stability.

These signals don't change during browser updates. They only change when:

• Hardware changes (new GPU)
• OS changes (new fonts installed)
• Locale changes

Best for: Long-term user identification across browser updates.

weightedId — Collision-Resistant

Each collector's contribution is weighted by its entropy. High-entropy collectors (canvas, WebGL, fonts) have more influence than low-entropy ones (screen, network).

This means that even if 100 corporate laptops have identical screen resolution, the fingerprints still differ because high-entropy signals dominate.

Best for: Environments with similar hardware.

crossBrowserId — Cross-Browser

Uses only hardware-level signals that are browser-independent, with normalization to absorb browser differences:

• GPU vendor/renderer — ANGLE strings stripped, model numbers removed, normalized to family name
• WebGL hardware params — GPU limits (maxRenderbuffer, maxUniformVectors, etc.)
• WebGPU limits — texture dimensions, buffer sizes (when available)
• Math precision — rounded to 8 significant digits (V8/SpiderMonkey/JSC tolerance)
• Screen resolution and DPR — colorDepth excluded (differs per browser)
• Timezone and locale — normalized to base language tag
• Installed fonts — browser-bundled fonts filtered out
• Audio sample rate

Best for: Identifying the same user across different browsers on the same device.

Choosing the Right ID

// One-time visitor identification
const visitorId = fp.id

// User recognition across browser updates
const userId = fp.stableId

// Enterprise fraud detection (similar hardware)
const deviceId = fp.weightedId

// "Is this the same person using Chrome and Firefox?"
const personId = fp.crossBrowserId