User agent parser.

The User-Agent header a browser, app or bot sends with every request, also readable in JavaScript as navigator.userAgent. It names the browser, its engine and the operating system, in a format shaped by thirty years of compatibility workarounds.

Compatibility. Sites once served their best pages to Netscape (Mozilla), so other browsers claimed to be Mozilla; Chrome added AppleWebKit and Safari for the pages written for Safari, and Edge adds Chrome. Parsers read the most specific token, such as Edg/ or SamsungBrowser/, to find the real one.

Chromium's successor to the string. Sec-CH-UA (the brands), Sec-CH-UA-Mobile and Sec-CH-UA-Platform are sent with every request; the full version, the platform's version, the model and the architecture only when a site asks with an Accept-CH header, or through navigator.userAgentData. Safari and Firefox don't send them.

Rarely. Strings are frozen, faked and change with every release, so a check written today misreads tomorrow's browser. Test for the feature instead (CSS @supports, 'IntersectionObserver' in window); keep the user agent for analytics, bot handling and working around a bug in one known version.

Not from the string, which anyone can send. Look up the visiting IP's host name (it should end in googlebot.com, google.com or googleusercontent.com), then look that name up and check it returns the same IP, or compare the IP with the ranges Google publishes. Bing, OpenAI, Anthropic and others publish their own.

Your own browser's string, parsed, beside what the page can read directly: your Client Hints where the browser has them, your screen and viewport sizes, pixel ratio, and media features such as color scheme, reduced motion, pointer and hover. Use it when you report a bug or check what CSS sees on a device.

Paste it into the field, or pick one of the samples: Safari on iPhone and iPad, Chrome on Android and Mac, Samsung Internet, Edge, Firefox, Googlebot, GPTBot, ClaudeBot, Slack's link preview and curl. Use mine puts your own back.

No. The string is parsed in the page with the Bowser library, loaded the first time the tool runs, and the device facts are read from your browser's own APIs. Nothing is sent anywhere.

No. The User agent parser is free and opens without a sign-in, an account or an install.

Nothing. The User agent parser holds your input in the page's memory, so closing or reloading the tab clears it, and it saves nothing in this browser.

Bowser, an open source parser loaded the first time the tool runs, plus a list of 72 crawlers, AI agents, link previewers, monitors and HTTP clients by their published tokens. It also fixes what Bowser leaves out: Blink's version from the Chrome token, Gecko's from rv:, and the model of an Android phone.

Not from a string: Windows 10 and 11 both send Windows NT 10.0. For your own browser, if it's Chromium-based, the client hints panel reads the platform version and calls 13 or higher Windows 11.

One at a time: paste a line from the log into the field and it's parsed as you type. Anything with bot, crawler or spider in it that isn't on the list is still flagged as an unknown bot, with the token it found.

Yes. Parsing, the screen and the media features work in every browser; the client hints panel says when a browser, such as Safari or Firefox, doesn't expose them.

Closest to the User agent parser: URL parser, UTM builder, HTTP status codes, JavaScript key code finder. Each opens on its own page, free and in the browser.

The User agent parser is one of 14 in its group. The group's page, /tools/web, lists them all.

The tool is free. Its interface is built from shadcn.io's own components (Button, Input, Slider and the rest of the registry under /ui), which come in React, Vue, Svelte and Solid. Pro adds the blocks and templates for the product around it.