My Browser and System

• Browser, version and rendering engine.
• Operating system, version and architecture.
• Device type, vendor and model, where reported.
• Preferred language, screen resolution, window size and pixel ratio.
• Hardware details: available cores, estimated memory, touch support and graphics card via WebGL.

Most of it comes from the user agent, a string the browser sends with every request. The rest comes from properties the page itself can read, such as screen dimensions and core count.

None of it is obtained by intrusion: these are data any site reads automatically on load. That is precisely the point of the tool — showing what you already hand over on every visit.

The check is therefore local. Nothing is sent to a server or stored.

For historical compatibility reasons, virtually every browser also identifies itself as several others. The string commonly contains Mozilla, Safari and Chrome at once, even in browsers that are none of the three.

The origin is old: sites began serving different content per browser, and browsers began disguising themselves to receive the good version. The result is a field nobody can clean up without breaking the web.

On top of that, modern browsers are deliberately reducing user agent detail to hinder tracking, which makes version detection steadily less reliable.

The practical conclusion for developers: **detect features, not browsers**. Checking whether a capability exists is robust; inferring it from a browser name breaks with every update.

The combination of all this data — fonts, resolution, graphics card, time zone, language, cores — forms an identifier that is usually near-unique, even without cookies.

This is the technique known as browser fingerprinting, used for persistent tracking precisely because it depends on nothing you can delete.

Seeing the full list here is a direct way to understand the scale of that exposure.

• Operating system name and version.
• Processor architecture, where reported.
• The platform the browser declares.
• Device type and whether touch is supported.

A web page has no access to your operating system. What it reads is whatever the browser chooses to declare — mainly through the user agent.

That means detection is always indirect and depends on the browser reporting honestly.

It is also why the displayed version is often approximate or generic, especially on newer systems.

Modern browsers deliberately freeze part of the user agent to hinder tracking. The result is that different systems declare the same generic version.

On Windows, recent versions frequently report as the previous one, because the internal numbering did not change the way sites expected. That is why a Windows 11 machine can be identified as Windows 10.

Something similar happens on macOS: the reported version tends to stick at an older number, also by privacy decision.

On iPad, the browser may declare itself as a desktop by default, which makes detection point at macOS rather than iPadOS.

If your site needs to know whether a capability exists, check the capability directly rather than inferring it from the operating system. Feature detection is reliable; system detection is not.

For installers and downloads, offering detection as a suggestion with a manual override is more robust than forcing the choice.

The reading happens in your browser. Nothing is sent to a server or stored.