What Is a User Agent?
A user agent is the software that acts on your behalf on the web, usually your browser, and its user-agent string is the line of text it sends in the User-Agent HTTP header to tell a website which browser, version and operating system is asking, such as Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/156.0.0.0 Safari/537.36 for Chrome 156 on Windows (October 2026). Sites use it to serve the right version of a page, count their visitors and spot bots, which is why scrapers set it with care. The user agent says which browser is asking and the IP says from where: ProxyField residential proxies give you 412M+ residential IPs from $1.95/GB, and you can start with the Easy Trial: 1 GB for $2.95.
- residential IPs
- 412M+
- from, traffic never expires
- $1.95/GB
- Easy Trial
- 1 GB for $2.95
On this page 10 sections
- What does "user agent" mean?
- How to read a user-agent string
- Why every browser says Mozilla
- User-agent reduction and client hints
- How to check your user agent
- What websites use the user agent for
- What is a user agent in web scraping?
- User agent vs browser fingerprint
- The other "user agent": user agent stylesheets
- What Is a User Agent? FAQ
What does "user agent" mean?
In web standards, the user agent is the client program that makes requests for a user: a browser, but also a crawler, a command-line tool such as curl, or an app. The HTTP standard defines the User-Agent header as information "about the user agent originating the request", which servers use to identify interoperability problems, tailor responses and analyze traffic (RFC 9110, section 10.1.5). In everyday use, "user agent" usually means the string itself.
Every request carries it: the page, its images and scripts, and the API calls the page makes. Browsers also expose it to JavaScript as navigator.userAgent.
How to read a user-agent string
Take Chrome 156 on Windows (the October 2026 version) and split it into its parts:
| Part | Meaning |
|---|---|
Mozilla/5.0 | A compatibility token almost every browser sends, for historical reasons |
( | The platform: Windows (also on Windows 11), 64-bit |
AppleWebKit/537.36 ( | The rendering engine family, kept for compatibility; Chrome's engine is Blink |
Chrome/156.0.0.0 | The browser and its major version; the rest is zeroed |
Safari/537.36 | Another compatibility token |
Other browsers follow the same pattern with their own tokens: Firefox adds rv:<version> and Gecko/20100101 Firefox/<version>, Edge appends Edg/<version>, and Safari sends Version/<version> Safari/605.1.15. MDN's User-Agent reference documents the formats. Current strings for each browser and device are in our list of user agents for scraping.
Why every browser says Mozilla
In the 1990s, sites sent their best pages only to Netscape, whose user agent began with Mozilla. Other browsers started sending Mozilla too, so they would get the same pages, and each new browser copied the tokens of the one before it to avoid being treated as unknown. That is why Chrome's string still names Mozilla, AppleWebKit, KHTML, Gecko and Safari, and why the string is a poor way to identify a browser exactly.
User-agent reduction and client hints
Because the full string helped sites fingerprint users, Chromium browsers now send a reduced one: only the major version (156.0.0.0), a fixed Windows NT 10.0 or Intel Mac OS X 10_15_7, and on Android a placeholder Android 10; K instead of the real version and device model. The details moved into client hints, separate headers such as Sec-CH-UA, Sec-CH-UA-Mobile and Sec-CH-UA-Platform, and a site can ask for more, such as the device model, with Accept-CH (MDN: Client hints). Safari and Firefox have frozen parts of their strings as well.
For a site, this means the User-Agent header alone says less than it used to. For a scraper, it means a Chrome user agent should come with Chrome's client hints, or it contradicts itself.
How to check your user agent
Three quick ways:
-
In the browser
Open DevTools, go to the Console and type
navigator.userAgent. -
In DevTools Network
Click any request and read
User-Agentunder Request Headers, together with the client hints. -
For a script or proxy
Request an echo endpoint such as
https://httpbin.org/user-agentfrom the same tool, so you see exactly what a site receives:
curl https://httpbin.org/user-agent
# {"user-agent": "curl/<version>"}
curl -x gate.proxyfield.io:7000 -U "<login>-country-us-city-new_york:<password>" -A "MyCrawler/1.0 (+https://example.com/bot)" https://httpbin.org/user-agentTo see the IP and location a site sees through a proxy, use our proxy checker.
What websites use the user agent for
- Serving the right page: a mobile layout for phones, a download link for your operating system, a fallback for an old browser.
- Analytics: browser and device shares in traffic reports.
- Fixing bugs: working around a problem in one browser version, the use RFC 9110 names first.
- Security and bot detection: blocking clients that announce themselves as scripts, such as
python-requestsorcurl, and comparing the claimed browser with what the request and the page show.
What is a user agent in web scraping?
In web scraping, the user agent is the string your scraper sends to say what it is. HTTP libraries send their own name by default (python-requests/<version>, curl/<version>, Scrapy/<version>), and many sites block those strings or serve them a different page. That is why scrapers replace it with a real browser's string, or, for an open crawler, with an honest name and a contact URL.
A browser string is a claim, though, and sites check it. A Chrome user agent with no client hints, a Python TLS handshake or a Linux font list is a contradiction that anti-bot systems look for. How to send a complete, consistent set is in HTTP headers for web scraping, and what the page itself reveals in browser fingerprinting in web scraping.
User agent vs browser fingerprint
The user agent is one value the browser states about itself. A browser fingerprint is the combination of many values the site measures, including the user agent, screen size, fonts, time zone, canvas and WebGL output. You can change the user agent with one line; the fingerprint has to agree with it, or the change makes the client more unusual, not less. The full picture is in what is a browser fingerprint.
The other "user agent": user agent stylesheets
If you met the term in CSS, it means something related but different: the user agent stylesheet is the browser's built-in default CSS, which gives headings their size and links their blue color before a site's own styles apply. DevTools labels those rules "user agent stylesheet". It has nothing to do with the User-Agent header except that both belong to the browser.
A browser string tells a site what you are; the IP tells it where you are. ProxyField residential proxies: IPs from real home internet providers with country, region and city targeting, from $1.95/GB with no subscription. Start with the Easy Trial (1 GB); traffic never expires, and all purchases are final.
What Is a User Agent? FAQ
What is a user agent?
ProxyField's 412M+ residential IPs say where a request comes from, and the user agent says which software sends it: a user agent is the program that makes web requests for you, usually your browser, and the user-agent string is the line it sends in the User-Agent header with its name, version and operating system, such as Chrome/156.0.0.0 on Windows NT 10.0. Sites read it to serve the right page and to spot bots.
What does user agent mean?
It means the client software acting for a user: a browser, a crawler, an app or a tool such as curl. In everyday use it also means the string that software sends to identify itself.
What is a user agent string?
It is the text value of the User-Agent header, made of tokens such as Mozilla/5.0, the platform in brackets, the engine and the browser with its version. Browsers expose the same value to JavaScript as navigator.userAgent.
What is an example of a user agent?
Firefox 158 on Windows (October 2026) sends Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:158.0) Gecko/20100101 Firefox/158.0, and curl sends curl/ followed by its version. Search engine crawlers send strings such as Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html).
What does the user agent tell a website?
The browser family and major version, the operating system and whether the device is mobile. Since user-agent reduction, Chromium browsers leave out the exact version and device model, which sites can request separately through client hints.
What is a user agent in web scraping?
It is the string your scraper sends to say what client it is. Libraries send their own name by default, such as python-requests/<version>, so scrapers replace it with a current browser string, or an honest crawler name with a contact URL.
Why are user agents important for web scraping?
Because the user agent is one of the first things a site checks: library defaults are often blocked outright or served a different page, and a browser string that contradicts the rest of the request is a bot signal. A current, consistent user agent avoids the easiest blocks; it doesn't fix the IP or the fingerprint.
Can you fake a user agent?
Yes, any client can send any string, and browsers let you override it in DevTools or with an extension. But it is only a claim: sites compare it with the client hints, the TLS handshake and what the page measures, so a fake that doesn't match the rest is easy to spot.
Does changing the user agent prevent anti-bot detection?
Not on its own. It gets you past rules that block library names such as python-requests, but anti-bot systems also check the IP, the other headers, the TLS fingerprint and the browser's behavior. Change it as part of a consistent setup, not as the only fix.
Is a user agent the same as a browser fingerprint?
No. The user agent is one value the browser states; a fingerprint combines many measured values, such as screen size, fonts, time zone, canvas and WebGL, with the user agent among them. Changing the user agent without the rest makes a fingerprint more unusual.
Show 2 more questionsShow fewer questions
Why does every user agent start with Mozilla/5.0?
For compatibility: early sites served their best pages only to Netscape, whose string began with Mozilla, so other browsers copied the token, and each new browser kept it. It no longer means anything about the browser.
What is a user agent stylesheet?
It is the browser's built-in default CSS, which styles headings, links and form fields before a site's own CSS applies; DevTools labels those rules "user agent stylesheet". It is unrelated to the User-Agent header except that both come from the browser.
Related pages
Easy Trial: 1 GB for $2.95.
Try for the price of 1 GB, no subscription, traffic never expires.