Cover Your Tracks (EFF)

The EFF‘s browser-fingerprinting test, and the longest-running instrument in this corpus. Launched 2010 as Panopticlick, the data-collection site behind eckersley-browser-uniqueness; rebranded and relaunched as Cover Your Tracks in 2020, now aimed at tracker protection as well as uniqueness. T1 — first-party documentation from the organisation that runs it, read 2026-08-13.

What it measures now

The 2010 list plus the things that did not exist then: user agent and HTTP headers, screen resolution and colour depth, timezone, system language, plugins, system fonts, whether JavaScript runs, cookie behaviour, platform and touch support, the Do Not Track header — and a hash of the image produced by canvas fingerprinting, plus WebGL hashes.

The canvas and WebGL additions are the sixteen-year delta, and they line up with what laperdrix-amiunique measured in 2016: the discriminating power moved to APIs that reach through the browser to the GPU and driver.

What it gives the visitor

A uniqueness score — how identifiable this browser is against everyone else who has taken the test — alongside a report on whether the browser blocks tracking ads and invisible trackers. It is the same self-audit posture myip takes for the network layer, run by an advocacy organisation rather than a developer, and it is the tool nothing-private points at without using.

The stance, stated plainly

EFF’s own account is not reassuring, and the honesty is the reason to hold this page:

• Do Not Track is ignored by “most websites and online trackers — with some laudable exceptions.”
• The practical recommendations are Privacy Badger, Disconnect or the Tor Browser, with the explicit caveat that against fingerprinting “none of them are ideal.”

Sixteen years after the organisation published the first measurement of the problem, its own advice is that the available defences are partial. Set that beside nothing-private‘s countermeasure table, which aims at a single value on a single API, and the corpus has two independent statements that the user-side defence is thin.

Caveats

• A live product, not a dataset. What it measures changes as the web does; this page is a 2026-08-13 snapshot of the about/learn pages, not of the code.
• Self-selected sample, permanently. Its uniqueness score compares you against people who chose to visit a fingerprinting test — the same bias eckersley-browser-uniqueness declared in 2010, now structural to the instrument.
• The page was read, not run. No test was executed from this host, so nothing here reports what it would say about the machine this wiki runs on.

Related

browser-fingerprinting · eckersley-browser-uniqueness · laperdrix-amiunique · nothing-private · myip · electronic-frontier-foundation · synthesis