User-Agent Parser
Browser-Strings dekodieren
| Browser | - |
| Engine | - |
| OS | - |
| Device | (desktop) |
| CPU | - |
Über dieses Tool
Decode any User-Agent string into structured fields - browser, engine, OS, device and CPU. Useful for debugging analytics, building feature-detection logic, identifying mystery traffic in server logs, or just figuring out what a weird UA actually represents.
Privvert uses ua-parser-js, the same library many web analytics tools use. Sample strings for iPhone, Android, Windows, macOS, curl, and common bots are one click away. Runs locally.
Funktionen
- Parses browser, engine, OS, device, CPU
- Sample strings for iPhone, Android, Windows, macOS, curl, common bots
- Loads your current browser's UA with one click
- Powered by ua-parser-js
- Browser-only - UA strings never uploaded
- Free and unlimited
- Breaks down the User-Agent into browser, engine, OS and device
So funktioniert's
- Paste a UA string, or click 'Use my browser' to load yours.
- Read the parsed fields.
Alles passiert direkt in deinem Browser mit JavaScript und WebAssembly. Deine Dateien werden nie hochgeladen, nie gespeichert und nie von uns gesehen.
Häufig gestellte Fragen
curl isn't a browser - its User-Agent only identifies itself ('curl/8.0.1'), so the browser/engine fields stay empty. That's correct behavior.
No - they're easy to fake and many privacy-focused browsers reduce or randomise them. Use User-Agent for analytics aggregates, never for security decisions.
Historical accident - Netscape was 'Mozilla', then every later browser claimed Mozilla compatibility for compatibility detection. The 'Mozilla/5.0' prefix is now a vestigial header.
Modern browsers are moving toward Client Hints (Sec-CH-UA) which give explicit, structured information instead of UA-string parsing. UA strings still exist as a fallback.
Historical accidents. Every new browser has copied parts of the older browsers' strings to convince servers that they're compatible, so the result is a string like Mozilla/5.0 (... AppleWebKit/...) ... Chrome/... Safari/... that contains the names of four browsers even though only one is actually running.
Almost never. Use feature detection (testing for the actual API you need) instead of guessing capabilities from the user-agent. Modern browsers also implement client hints, which give you cleaner structured data when you do need to identify the client.