What is a proxy? How proxies work, in plain English

Every request you make on the internet carries a return address: your IP. Websites read it, remember it, and decide how to treat you based on it. A proxy puts a different return address on the envelope. That is the whole idea. Everything else in this industry is detail on top of it.

The one-sentence version

A proxy is a server that sits between you and the site you are talking to. You send your request to the proxy, the proxy sends it on to the site, the site answers the proxy, and the proxy hands the answer back to you. The site only ever sees the proxy's IP address, never yours.

That is different from a VPN, which wraps your whole device's traffic in an encrypted tunnel. A proxy is applied per connection or per application, which is exactly what you want when a script needs to make ten thousand requests that each look like they came from somewhere different.

Why anyone needs one

Websites do not treat all visitors the same. Almost every large site keeps a running score per IP address: how many requests it has made, how fast, to which pages, with which headers. Cross a threshold and the site slows you down, shows a captcha, or blocks the address outright. This is rate limiting, and it is the reason the proxy industry exists.

A few real situations:

• A price-comparison tool needs today's prices for 40,000 products across six retailers. From one IP, retailer number one blocks it after a few hundred pages.
• An ad-verification team needs to see the ads actually served to a user in Madrid, on a phone, on a Spanish carrier. From an office in Berlin, they see the Berlin version.
• A brand-protection team searches marketplaces for counterfeits. Marketplaces show a sanitised catalogue to anything that looks like a bot.
• A developer tests how their own site behaves for visitors in Brazil before launch.

In each case the fix is the same: make the requests from many different, ordinary-looking addresses in the right places. That is what a proxy network provides.

How a request actually flows

Say your script wants the page at https://example.com/products. Configured to use a Nyxon proxy, it does this instead:

your script  ->  gateway (gw.nyxon.sh)  ->  exit IP in the target country  ->  example.com
your script  <-  gateway                <-  exit IP                         <-  example.com

The gateway authenticates you with a username and password, reads the targeting options in your username (country, city, session), picks an exit that matches, and relays the traffic. The site sees a request from a residential connection in, say, Dallas. When the response comes back the gateway returns it to you unchanged.

Rotation is what makes this scale. If every request goes out through a different exit, no single IP ever accumulates enough history to trip a rate limit. If you need the same IP for a while, for a login session or a shopping cart, you ask for a sticky session instead and the gateway keeps routing you through the same exit.

Where do the IPs come from?

This is the question that separates providers, and it is worth understanding before you buy.

• Datacenter IPs belong to hosting companies. They are fast and cheap, and every website knows they are not real people. Fine for targets that do not check.
• ISP IPs are addresses registered to a consumer internet provider but hosted on datacenter hardware. They look residential to a site and stay static, which suits logins and long sessions.
• Residential IPs belong to real home connections. Traffic from them looks like a person on their sofa, because it is routed through one.
• Mobile IPs belong to carrier networks. Thousands of real phones share each address behind carrier-grade NAT, so blocking one blocks all of them. Sites almost never do.

Nyxon sells all four. The residential and mobile exits come from tier-1 carrier peering and are never resold second-hand lists, which is the part most budget providers get wrong. A pool that has been resold ten times has been burned on every major site ten times.

What a proxy does not do

A proxy changes your network identity, not your browser identity. If your script announces itself with a default python-requests user agent, a residential IP will not save it. Serious targets look at the TLS handshake, the browser fingerprint and the behaviour of the session as well as the IP. The proxy handles one layer. The rest is on your tooling, and we cover it in 10 proxy mistakes that get accounts banned.

A proxy also does not make anything legal that was not legal before. Read a site's terms, respect robots directives where they apply to you, and keep the request rate humane. Our own acceptable-use policy is short and enforced.

What it costs

Pricing in this industry is usually per gigabyte of traffic through the proxy. At Nyxon, residential starts at $0.20 per GB, mobile 4G/5G at $1.25 per GB, static ISP at $0.50 per GB and datacenter at $0.36 per GB. There is no subscription and no minimum; you top up a balance from $1 in crypto and it never expires. Every new account gets a free residential trial to test on real targets before spending anything.

Try it in sixty seconds

1. Create an account. No card, no identity check.
2. Claim the free trial from the dashboard.
3. Open Generate, pick a country and copy the proxy string.
4. Run it:

curl -x http://USERNAME:[email protected]:8080 https://api.ipify.org

The IP that comes back is the exit's, not yours. That is a proxy working.