An October 2026 comparison of the leading browser fingerprinting libraries: FingerprintJS, ThumbmarkJS, ClientJS, CreepJS, BotD, and Benny the Doorman, across license, maintenance, pricing, cross-browser identity, and anti-spoof. Here is how to choose.
A browser fingerprinting library reads signals from the browser, things like the canvas rendering, the audio stack, installed fonts, and the GPU, and turns them into a stable identifier you can use to recognize a returning visitor or device with no cookie required. The libraries below all run client-side in JavaScript and return that identifier as a hash.
Where they differ is licensing, how actively they are maintained, whether they recognize the same device across different browsers, whether they detect spoofing, and how much you have to build yourself. This guide compares the six that matter as of October 2026, then gives a plain rule of thumb for picking one.
What to look for in a fingerprinting library
- Licensing and source availability: MIT, Apache, source-available, or closed. This matters for audits, internal security review, and the ability to fork.
- Maintenance: browsers change their APIs and privacy defaults every few months. A library without a release in years will slowly lose accuracy.
- Cross-browser identity: does the same physical device produce a stable ID in Chrome, Safari, and Firefox, or a different hash in each browser?
- Comparison logic: does it ship an API to score how similar two fingerprints are, or do you write that matching code yourself?
- Anti-spoof: does it flag automation, anti-detect browsers, and tampered signals, and is that free or gated to a paid tier?
- Dependencies and runtime: client-only, or does the useful part need a server roundtrip, an API key, and per-call pricing?
Browser fingerprinting libraries at a glance (October 2026)
| Library | License | Latest release | Output | Cross-browser ID | Anti-spoof | Best for |
|---|---|---|---|---|---|---|
| FingerprintJS (OSS) | MIT | v5.3.0, Oct 2026 | Single hash | No | No | Widest open-source signal set, drop-in |
| Fingerprint Pro | Commercial | Hosted API | Server visitor ID | Yes (server) | Yes (paid) | Enterprise device intelligence |
| ThumbmarkJS | MIT (paid API tier) | v1.12.0, Sep 2026 | Single hash; visitor ID on the paid API | No (OSS) | Paid API | Simple drop-in with a low-cost hosted tier |
| ClientJS | Apache 2.0 | v0.2.1, Oct 2021 | Single hash | No | No | Lightweight, legacy projects |
| CreepJS | MIT | No versioned releases | Detailed report + trust score | No | Yes (detection-focused) | Research, auditing your own entropy |
| BotD | MIT | v2.0.0, Nov 2025 | Bot / not-bot verdict | Not applicable | Automation only | Free client-side bot detection |
| Benny the Doorman | Free (closed source) | Actively released | Two hashes (per-browser + hardware) | Yes (client-side) | Yes (free) | Cross-browser device matching + free anti-spoof |
FingerprintJS
FingerprintJS is the most widely known name in browser fingerprinting. The open-source library is MIT-licensed, actively maintained (v5.3.0 shipped in October 2026), and produces a single hash from a large signal set. It is a solid drop-in for recognizing a returning visitor inside the same browser.
Its own README is candid about the limits: it says the open-source version's accuracy is significantly lower than the commercial product, and that browser-side fingerprints are vulnerable to spoofing and reverse engineering. Server-resolved persistent IDs, tamper detection, and IP intelligence live in Fingerprint Pro, a hosted platform rather than a library. Pro has a free plan of 1,000 API calls per month, then Pro Plus at $99 per month for 20,000 calls and $4 per additional 1,000.
If you need that platform and have the budget, Pro is the most mature option. If you only need a client-side hash, the breadth of the open-source collector set is its main draw. The [Benny vs FingerprintJS comparison](/vs/fingerprintjs) breaks down the OSS and Pro tiers in detail, and [FingerprintJS alternatives](/blog/fingerprintjs-alternatives) covers eight options side by side.
ThumbmarkJS
ThumbmarkJS is a modern, MIT-licensed library with a steady, roughly monthly release cadence (v1.12.0 shipped in September 2026). The open-source version produces a single hash and is genuinely simple to drop in.
It now has a hosted API alongside the library. ThumbmarkJS states the open-source fingerprint reaches about 80% uniqueness and the API version about 99%, because the API adds server-side signals such as TLS and HTTP header characteristics and returns a server-generated visitor ID. The API has a free allowance of 1,000 calls per month, and the Pro plan is EUR 15 per month for 15,000 calls, which makes it the cheapest hosted tier in this list.
The open-source library does not split hardware from engine signals, so it does not give you a deterministic cross-browser device ID on its own. For a single-hash model with an MIT license and an inexpensive upgrade path, it is a strong pick. The [Benny vs ThumbmarkJS comparison](/vs/thumbmarkjs) covers the specifics.
ClientJS
ClientJS is one of the older open-source fingerprinting libraries, Apache-2.0-licensed and lightweight. It wraps a set of browser properties into a single numeric fingerprint and is easy to read and integrate.
Its last release, v0.2.1, was in October 2021, so it has not kept pace with newer signal techniques or with four years of browser privacy changes. It has no cross-browser identity or anti-spoof. It is a reasonable choice for simple, low-stakes recognition in a legacy codebase, but new projects are better served by a maintained library. See [Benny vs ClientJS](/vs/clientjs) for a side-by-side.
CreepJS
CreepJS is different from the others: it is an MIT-licensed research and demonstration project, not a production identification library. It collects an unusually deep set of signals and focuses on detecting lies, such as spoofed user agents, anti-detect browsers, and tampered APIs, and presents a detailed trust report.
Developers use it to understand how identifiable their own browser is and to study anti-detect techniques. It has no versioned npm releases and is not designed to be dropped into an app as a stable device ID. Its README now warns that the only official deployment is abrahamjuliot.github.io/creepjs and that copies on other domains should be treated as untrusted. More in [CreepJS explained](/vs/creepjs).
BotD
BotD is FingerprintJS's free, MIT-licensed bot detection library. It does not identify visitors at all; it runs client-side and returns a verdict on whether the browser is being driven by automation tooling.
It is a useful companion to any identification library in this list when your problem is scripted traffic rather than returning humans. It is not a substitute for a device ID, and like any client-side check it can be studied and evaded by a determined operator.
Benny the Doorman
Benny the Doorman is the library this site maintains, so treat this entry as the vendor's own view. Its distinguishing feature is that it tags every signal as hardware-bound or engine-bound and returns two hashes: a per-browser fingerprint, and a hardwareFingerprint that stays identical across Chrome, Safari, Firefox, and Brave on the same device. [Hardware vs engine fingerprints](/blog/hardware-vs-engine-fingerprint) explains why that split matters.
It ships a compareFingerprints API with [six match modes](/blog/compare-device-fingerprints-match-modes) and a rule-based anti-spoof consistency score on every result, all in the free tier and all client-side. It has a narrower raw signal count than FingerprintJS and is closed-source, which are real trade-offs. It is the cleanest fit when you need cross-browser device matching or free anti-spoof without standing up a server.
Beyond JavaScript: server-side and non-browser options
Every library above runs in the browser. If you are asking which Python or backend library to use, the answer is a different category: server-side fingerprinting works from what reaches your server, such as the TLS handshake and HTTP headers, rather than from JavaScript APIs. Techniques like [JA3 and JA4 TLS fingerprinting](/glossary/ja3-fingerprinting) are implemented in proxies, CDNs, and network tooling rather than in an npm package.
Server-side signals are harder to tamper with from the page, but they identify a client stack, not a device, and they collapse many users behind the same proxy or browser version. Most production setups combine both layers, which is also why the hosted tiers of Fingerprint Pro and ThumbmarkJS add server-side signals. [Server-side fingerprinting](/glossary/server-side-fingerprinting) covers the trade-offs.
How to choose
Match the tool to the job. If you need an enterprise platform with server-resolved identity, IP intelligence, and native mobile SDKs, Fingerprint Pro is built for that. If you need a simple per-browser hash with an MIT license, FingerprintJS OSS or ThumbmarkJS will do it, and ThumbmarkJS has the cheapest hosted upgrade. If you are auditing your own fingerprint or studying anti-detect browsers, CreepJS is the research tool. If the problem is bots rather than people, add BotD.
If your problem is recognizing the same device across browsers, the common case in anti-fraud, trial-abuse prevention, and duplicate-account detection, you need hardware-bound identity, which is where Benny's two-hash model is aimed.
Most teams overestimate how much signal breadth they need and underestimate how often the same user returns in a different browser. A single per-browser hash reads that user as two devices. If that failure mode matters for your use case, prioritize cross-browser identity over raw collector count.
Head-to-head comparisons
- [Benny vs FingerprintJS](/vs/fingerprintjs): open-source and Pro tiers, pricing, and cross-browser identity.
- [Benny vs ThumbmarkJS](/vs/thumbmarkjs): single hash vs two-hash model.
- [Benny vs ClientJS](/vs/clientjs): the classic library vs a maintained rebuild.
- [CreepJS explained](/vs/creepjs): research tool vs production library.
- [Benny vs SEON](/vs/seon): a library vs a full anti-fraud platform.
- [Open-source fingerprinting libraries, compared](/blog/open-source-fingerprinting-libraries): the license-first cut of this list.
Frequently asked questions
What is the best free browser fingerprinting library?
It depends on the job. For a simple open-source per-browser hash, FingerprintJS OSS and ThumbmarkJS are the strongest maintained options, both MIT-licensed. For cross-browser device identity and free anti-spoof without a server, Benny the Doorman is purpose-built, though it is closed-source.
What is the best open-source alternative to FingerprintJS?
ThumbmarkJS is the closest modern equivalent: MIT-licensed, actively maintained, and simple to drop in, with an inexpensive hosted API if you later need server-side signals. ClientJS is a lighter option but has not had a release since 2021. If your real need is cross-browser device matching, which FingerprintJS only offers through its paid Pro platform, Benny the Doorman provides a cross-browser hardware ID for free.
Is FingerprintJS free?
The FingerprintJS open-source library is free and MIT-licensed, and returns a single per-browser hash. The commercial Fingerprint Pro platform has a free plan of 1,000 API calls per month, then Pro Plus at 99 dollars per month for 20,000 calls and 4 dollars per additional 1,000, with custom Enterprise pricing above that (as published in October 2026).
Is ThumbmarkJS free?
The ThumbmarkJS library is free and MIT-licensed. Its hosted API, which adds server-side signals and a server-generated visitor ID, has a free allowance of 1,000 calls per month and a Pro plan at 15 euros per month for 15,000 calls (as published in October 2026).
Do browser fingerprinting libraries work in incognito or private mode?
Yes. Fingerprinting reads device and browser characteristics rather than stored cookies, so most signals are still available in incognito or private mode. Hardware-bound signals in particular stay the same between a normal and a private session on the same device, which is why a fingerprint survives where a cookie does not.
Is there a Python browser fingerprinting library?
Browser fingerprinting has to run in the browser, so the libraries are JavaScript. On the server, the equivalent is server-side fingerprinting from the TLS handshake and HTTP headers, such as JA3 or JA4, which is usually done in a proxy, CDN, or network tool rather than a Python package. Most production setups combine a client-side library with server-side signals.
What is CreepJS used for?
CreepJS is a research and demonstration tool that detects fingerprint spoofing and shows how identifiable a browser is. It is meant for auditing your own entropy and studying anti-detect browsers, not for use as a production device identifier. For a stable ID in your application, a library like FingerprintJS, ThumbmarkJS, or Benny the Doorman is the right category.
Are browser fingerprinting libraries legal to use?
Fingerprinting is legal in most jurisdictions but regulated. The EU ePrivacy Directive and GDPR, the UK GDPR, and various national and US state laws govern it, and many include an exemption for fraud prevention and security. Legality depends on your purpose, your disclosure, and your region. This is engineering guidance, not legal advice.
Get started
Need cross-browser device identity?
Benny the Doorman returns a per-browser fingerprint and a deterministic cross-browser hardware ID in one client-side call, with free anti-spoof scoring. One npm install, no account, no cookies.
Last reviewed October 23, 2026

