The User-Agent parser reads this browser's navigator.userAgent by default, and it will just as happily take a string you paste in. An in-app copy of ua-parser-js then reports the browser, rendering engine, operating system, device and CPU architecture. Everything is an inference from string rules — no remote probe of real hardware takes place.

1. 01Keep the current UA, or paste in a User-Agent string.
2. 02Or click a sample such as iPhone Safari, Android Chrome, Windows Edge or WeChat.
3. 03Read off the browser and version, the engine, the OS, the device type, vendor and model, and the CPU architecture.
4. 04Copy the raw UA when you need it elsewhere.

• What can be recognised depends on the rule set in the current ua-parser-js and on what the UA itself states.
• Privacy-reduced UAs, frozen UAs and spoofing leave version, model and CPU fields missing or wrong.
• When the device type is absent the page falls back to “desktop device (assumed)”; that is a UI fallback, not proof that the device really is a desktop.
• The tool does not read Client Hints headers from a server and does not query an online device database.
• A pasted UA is only parsed as a string: no browser is emulated and the current page's UA does not change.

The UA you enter and the parsed result are never uploaded; the default value is information the browser already exposes to the page anyway.