User Agent Parser
Identify browser, OS and device from a UA.
How to use User Agent Parser
Read a user-agent string and guess the browser family, operating system and device type.
Open User Agent Parser on toolpermarket — the form is ready, so type your numbers and read the result.
In the User agent field, paste a user-agent string captured from a request or log. Leave it exactly as captured, including the parentheses.
Press Parse to compute the result. The output appears immediately below the form. Edit a value and press the button once more to see the updated answer.
Read the result and use it as needed. If the number looks off, re-check your inputs and run it again — the math is deterministic, so the same entries always return the same answer.
Worked example. Paste "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0 Safari/537.36" to see Chrome on macOS, desktop.
Tips for accurate results.
The browser guess follows the last recognised engine token.
OS is inferred from the platform fragment inside parentheses.
Mobile vs desktop is decided by phone or tablet keywords.
When to use User Agent Parser. Quickly labelling traffic in a log review. Guessing a visitor platform before building a fix. Explaining a UA string to a teammate.
When you actually reach for this
- You need to know a visitor's browser, OS, and device from a raw User-Agent string for analytics or logging.
- You are reproducing a bug reported on "Safari on iOS 17" and want to confirm the UA really says that.
Where this tool stops being accurate
- User-Agent strings are freely spoofable, so never use them for security decisions — only for heuristics and stats.
- Modern UA reduction (Chrome's frozen UA) hides fine detail, so exact minor versions may no longer be available.
Frequently asked questions
How accurate is the guess?
It is a heuristic based on well-known tokens in the string, so it is right for common browsers and OSes but can mislabel unusual or spoofed user agents.
Can it detect bots or crawlers?
Only if the string contains a recognizable crawler token; many bots mimic browser UAs, so absence of a bot token does not prove a human visitor.
Why does it say "Unknown"?
When the string lacks the expected browser or OS keywords, the tool reports Unknown rather than guessing, so you are not misled by a malformed UA.
Why does the parser say "Unknown" for a real browser?
The UA may be from a bot, a privacy-hardened browser, or a reduced UA that omits the model. Treat unknown as "don't assume," not an error.
Can I block users based on User-Agent?
Not reliably — it is trivially faked. Use it for analytics only; enforce access with authentication and server-side checks.