User Agent Finder

User Agent Finder

Your browser’s user agent appears in User agent string as soon as the page opens, read and explained, with a Copy button.

To read a different one, paste it into the box: from a server log, an analytics report or a support ticket. Or pick one from Try an example: Chrome on Windows, Safari on iPhone, Firefox on Ubuntu, Samsung Internet on a Galaxy, Edge on a Mac, an iPad asking for desktop sites, Googlebot, GPTBot, curl and Internet Explorer 11. Mine brings back your own.

Example: Chrome on Linux

Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/155.0.0.0 Safari/537.36

The tool reads this as Chrome 155 on Linux, desktop or laptop, reduced user agent:

• Browser: Chrome, version 155.0.0.0. The full version, 155.0.8059.84, comes from client hints, not from the string.
• Engine: Blink.
• Operating system: Linux, with no version given.
• Device: desktop or laptop.

Then the string is taken apart piece by piece:

Almost every piece is there for history’s sake. In the 1990s, sites sent their best pages only to browsers that said “Mozilla”, so every newer browser claimed to be Mozilla too. Each new engine since then has added the names of the ones before it.

Why Chrome’s string is “reduced”

The user agent is sent with every request, and its details helped websites tell visitors apart. So Chrome and every browser built on it now send a fixed, reduced string. Only the major version is real, and the platform part is one of a few set values, whatever the actual system:

That explains some puzzling analytics reports. Every Windows 11 visitor shows up as Windows 10. Every Mac shows up as an Intel Mac running a 2019 version of macOS. And the phone model is gone from Android user agents.

Client hints: what Chrome tells only when asked

The details removed from the string are still available as client hints, extra headers a browser sends only when a site asks for them. For your own browser, the tool shows what yours reports:

• Brands: the browser names and full versions, here Google Chrome and Chromium 155.0.8059.84.
• Platform, Architecture (x86 64-bit) and Mobile (no).

Every Chrome-based browser sends three hints with each secure request: the brand list, whether it’s a mobile device, and the platform name. The rest (full version, platform version, architecture, device model) a site has to request with an Accept-CH response header, or in JavaScript with navigator.userAgentData.getHighEntropyValues(). That’s how a site can tell Windows 11 apart: it reports a platform version of 13 or higher.

Two details trip people up. The brand list also contains a deliberately meaningless entry such as Not_A Brand, added so that sites can’t rely on an exact list. And Firefox and Safari don’t send client hints at all: their user agent string is all there is.

Reading other user agents

The examples show strings worth knowing:

• An iPad asking for desktop sites sends a Mac’s user agent. Since iPadOS 13, Safari on an iPad does this by default, so iPads are hard to tell from Macs by user agent alone.
• Edge looks like Chrome plus an Edg/ token. Samsung Internet adds SamsungBrowser/. Other Chromium browsers each add their own token.
• Internet Explorer 11 never says “MSIE”. It identifies itself with Trident/7.0; rv:11.0, a deliberate change to stop sites from sending it old IE-only pages.
• curl sends just curl/ and a version. Most scripts and command-line tools are that easy to spot, unless they’ve been set to pretend otherwise.

Bots and what they’re for

The tool recognises ninety bots and says what each one is for:

• Search engines, like Googlebot and Bingbot.
• AI crawlers, like OpenAI’s GPTBot, which collect pages for AI models and AI search.
• SEO tools, which build backlink and ranking databases.
• Link previews, which fetch the title and image when someone shares your link in a chat or social app.
• Monitoring services, which check that your site is up.
• Command-line tools and code libraries.

This matters when you write robots.txt rules or firewall rules. Block an AI crawler and you only opt out of that use. Block link-preview bots and your links appear without a title or image when people share them. Some controls have no user agent of their own: Google-Extended, for instance, is only a robots.txt token that controls whether Googlebot’s crawl can be used for Google’s AI models.

Keep in mind that a user agent is only a claim. Any scraper can call itself Googlebot. To check a real one, look up the visitor’s IP address with IP to Hostname: a genuine Googlebot has a name ending in googlebot.com or google.com that points back to the same address.

What it’s useful for

• Support requests. Copy your user agent into a bug report so the developer knows exactly what you’re using. For more detail about your browser, screen and settings, use What’s My Browser.
• Reading server logs. Paste an unfamiliar string from your access log to see whether it’s a person, a crawler or a script, and what that crawler is for.
• Seeing your site as Googlebot. In Chrome DevTools, under Network conditions, you can set a custom user agent. Paste the Googlebot smartphone example and reload to see what your site serves it.
• Making sense of analytics. If your stats show no Windows 11 and only one macOS version, that’s the reduced user agent, not your audience.

FAQ

Can I change my user agent?

Yes. Chrome and Edge let you set one in DevTools under Network conditions, Firefox has a setting in about:config, and extensions can switch it for you. Command-line tools let you set any string (curl -A "..."). That’s why a user agent should never be trusted for security decisions.

Should my site check the user agent to see what a browser supports?

Rarely. Browsers copy each other’s tokens, and the details are now reduced, so checks based on the string break easily. Testing for the feature itself in JavaScript is more reliable. The user agent is still useful for statistics and for recognising bots.

Does my user agent reveal who I am?

Not on its own: no name, no location, nothing about your account. But it adds to the details that make a browser recognisable, which is why it was reduced. The Browser Fingerprint Audit shows how unique your browser looks as a whole.

Why does Googlebot’s user agent mention Chrome?

Googlebot renders pages with an up-to-date version of Chrome and says so in its user agent. Google’s documentation writes the version as W.X.Y.Z, because it changes as Chrome updates.