This page turns the clues from an ordinary page visit into a stranger’s first impression of you. It shows the useful guess and the reason that guess can fail.
Plain English translates your time, language, device, preferences, and behavior into the kinds of buckets a marketer or ad system uses. A guess is never presented as a proven fact.
“For Nerds” keeps the exact browser values, source APIs, how long each value tends to last, how useful it is for recognizing you, and the permission boundaries behind every plain-language conclusion.
To demonstrate return-visit recognition, this page saves one item to your browser’s local storage for this site: a short one-way summary of a few stable-looking browser values, plus a visit count and timestamps. No name, no account, nothing sent anywhere. The “Recognition on a later visit” row shows you the result, and Clear this site’s saved data at the bottom of the page erases it along with cookies, session storage, databases, and cached copies.
The report is assembled in your tab. None of the signal values are transmitted, stored, or compared against a database. The page does load Google Analytics for aggregate visit counts — and rather than hide that, the report tests for it and shows you the result. Look for “Google Analytics on this page.” That tag is the only third party this page loads: beyond it there are no ad SDKs, no other trackers, no external fonts, and no other third-party scripts.
A browser setting is not a personality. A device is not an income. A time zone is not a home address. Real marketing systems get their power by combining clues like these with accounts, purchases, tracking IDs, or purchased data.
It reads the signals your browser hands over on an ordinary page visit — over 120 of them — and translates each one into plain English: what a site can conclude, how it knew, what the conclusion can’t prove, and how to reveal less. Every browser check runs in your own tab.
Yes — completely free, with no account, no sign-in, no limits, and no paid tier. Nothing about your browser is uploaded, sold, or added to a fingerprint database. The one thing kept on your device is a short return-visit signature in this site’s local storage, which powers the recognition demo and which you can erase with one button.
Open the page. The checks run on their own and the results appear as a list — there is no button to press. Choose any row to read what that signal means and how to reduce it. Switch to “For Nerds” for the exact value, the API it came from, how long it lasts, and how useful it is for recognizing you. Copy or export the whole report as JSON from there.
The browser report is built entirely inside your tab and none of the signal values are transmitted anywhere. Two things are worth knowing. First, the page saves a short one-way return-visit signature to this site’s local storage so it can demonstrate recognition — “Clear this site’s saved data” erases it along with cookies, databases, and cached copies. Second, the page loads Google Analytics for aggregate visit counts, and rather than hide that, the report tests for it and shows you the result — look for the “Google Analytics on this page” row. If you want to see what a real analytics tag looks like from the inside, this page is a live example of one.
Fingerprinting is recognizing a browser by the combination of details it exposes — fonts, screen geometry, graphics stack, audio stack, time zone, language, hardware hints — instead of by a cookie. No single value identifies you. Combined, they can be distinctive enough to link visits together after cookies are cleared and in private windows.
Because an honest uniqueness score needs a database. EFF’s Cover Your Tracks and AmIUnique compare your fingerprint against a corpus of previously observed visitors, which is what makes their “1 in N browsers” number mean something. This page has no server and no corpus, so any score it invented would be a guess dressed up as math. It shows you the raw surface instead, and marks which signals are repeatable enough to matter. Run Cover Your Tracks alongside it for the population comparison.
On this site, yes. The page reads it from Cloudflare’s /cdn-cgi/trace endpoint — a same-origin request to the edge server that already received your visit — so no third-party lookup service gets told you were here. That also supplies the country the connection suggests, the serving data center, and your TLS and protocol details. On a local copy or a non-Cloudflare host, those rows read “not available” instead.
BrowserLeaks.com is a deep suite of separate technical tests — IP and WebRTC leaks, DNS, TLS, HTTP/2 and TCP/IP fingerprinting — aimed at people who already know what those are. This page covers the client-side signals only and spends its effort on translation instead: one plain-English consequence per signal, plus a technical view when you want the exact value. Different jobs. Use both.
Yes. The layout adapts to phone screens, and mobile browsers expose their own extra signals — touch points, orientation, motion sensors, device memory class — so the report is genuinely different from your desktop result. Comparing the two is worth doing.
Every row includes a specific step under “Reveal less.” The general moves: use a mainstream browser at default settings so you blend in rather than stand out; keep precise location, camera, microphone, clipboard, and sensor permissions denied until a feature truly needs them; clear site data between visits you don’t want linked; and consider a browser with built-in fingerprint randomization. Blocking everything can backfire — an unusual configuration is itself distinctive.