It lists what a web page can learn about your browser without asking you anything: no permission prompt appears, because nothing here needs one. The report is grouped into the browser and system (User-Agent string and what it says, client hints), the screen (size, device pixel ratio, colour depth and gamut, HDR), language and time (languages, time zone, number and date formatting), input (mouse, touch, hover), preferences (dark mode, reduced motion, contrast, forced colours), hardware hints (CPU cores, rounded device memory), storage and privacy signals (cookies, storage, quota, Do Not Track, Global Privacy Control), network hints, graphics (the WebGL renderer, usually your GPU's name) and a table of supported features, from WebP and AVIF to WebAssembly, WebCodecs, passkeys and CSS :has().
Copy it as text to paste into a bug report, or as JSON for a script.
Nothing is sent, and there is no IP lookup
Every value is read by this page, in your browser, and stays there. The page makes no network request to build the report. That is also why it does not show your IP address or location: a page cannot see its own visitor's public IP without asking a server, and asking a server would mean sending something. Sites that show your IP do exactly that.
Reading the User-Agent string
The User-Agent is a long string every browser sends with every request. The tool's own small parser reads the browser, engine, operating system and device class from it, and says what the string can no longer tell. Browsers now freeze parts of it to make tracking harder: Chrome and Edge report Windows 10 and 11 alike as "Windows NT 10.0", macOS as 10.15.7 whatever the real version, and Android as "Android 10; K" without the model. Safari on an iPad asks for desktop sites and claims to be a Mac. On iPhone and iPad, browsers have generally been built on WebKit whatever their name. Any script can also change the string, so treat it as a hint.
Chromium-based browsers additionally expose client hints: low-entropy brands and platform by default, and on request higher-detail values such as the full version list, platform version and CPU architecture. The tool reads both (the request needs no permission) and filters out the deliberately fake "Not A Brand" entry.
Fingerprinting: why ordinary details matter
No single value here identifies you. A 1920 × 1080 screen, an English language list or a Europe/Berlin time zone are shared by millions. Combined, though, a dozen ordinary values (screen and available area, device pixel ratio, time zone, languages, platform, core count, memory, touch points, GPU renderer, User-Agent) can make one browser stand out, and any script can read them silently. That is browser fingerprinting, and unlike a cookie it cannot be deleted.
The tool marks the values commonly used this way with ● and computes a hash of them on the page, to show what a combined fingerprint looks like. It is an illustration, not a uniqueness score: estimating how rare your combination is would require comparing it with many other browsers on a server, which this page deliberately does not do. Things that reduce fingerprinting: browsers with built-in protections that report standardised values, tracker blockers, and keeping the browser up to date and at default settings. For the theory, read what your browser tells every website. Hardware checks live in the screen test and keyboard tester.
