What Is a Mobile Proxy? How 4G and 5G Proxies Work

What a mobile proxy is

A proxy is a server that makes requests on your behalf. The website talks to the proxy, not to you, so it sees the proxy address. What makes a proxy "mobile" is where that address comes from. With a datacenter proxy it belongs to a hosting company. With a mobile proxy it belongs to a mobile carrier such as T-Mobile, Vodafone or Three, because the last hop is a real SIM device on the carrier network.

Three addresses are involved, and mixing them up causes most of the confusion around mobile proxies. The entry address is what you type into your browser or script: a host, a port and a login. The device address is the private address the carrier gives the SIM device. The exit address is the public address the carrier uses for that device on the internet. Only the exit address is what websites, fraud filters and IP databases see.

A mobile proxy is not a VPN app on your phone. A VPN moves all of a device's traffic into a tunnel that ends on the VPN provider's server, usually on a hosting network. A mobile proxy works the other way round: the tunnel ends on a phone line, so the exit is on the carrier network.

How a mobile proxy server works

The phrase "mobile proxy server" covers two machines. One is the gateway, an ordinary server with a public address that accepts your connections and checks your login. The other is the device: an LTE or 5G USB modem, a modem board or an Android phone with a SIM card, holding an open mobile data session. A tunnel joins the two, because the device cannot accept incoming connections itself. Carriers do not give it a reachable public address.

Follow one request. Your scraper asks the gateway, over HTTP or SOCKS5, to fetch a page. The gateway hands the request down the tunnel to the device. The device opens the connection to the website over its mobile data session. The traffic leaves through the cell site to the carrier core, where the packet gateway (the PGW in a 4G network, the UPF in 5G) routes it to the internet through the carrier NAT. The answer comes back the same way. To the website, the visitor is a subscriber of that carrier.

This is why the entry address of a mobile proxy often looks like a datacenter in an IP lookup while the exit address does not. If you test the wrong one, the line looks fake when it is not. Why a mobile proxy shows as datacenter walks through that case.

CGNAT: why many phones share one address

There are not enough IPv4 addresses for every phone, so carriers share them. Each device gets a private address, often from 100.64.0.0/10, a block set aside for exactly this purpose by RFC 6598. A carrier-grade NAT translates those private addresses to a much smaller set of public ones, keeping subscribers apart by port. Many carriers hand each subscriber a block of ports on a shared address.

For a website this has a practical consequence. The same public address can stand for many real people over a day. Blocking it outright would lock out customers, so sites usually treat carrier addresses with more caution before blocking: they rate-limit, ask for a captcha or look harder at the account and browser. That caution is the main reason people use mobile proxies.

IPv6 works differently. A 3GPP network gives each data connection its own /64 prefix (RFC 6459), so an IPv6 address points to one connection, not a crowd. Some carriers now run IPv6-only networks and reach IPv4 sites through translation such as 464XLAT (RFC 6877), which puts the IPv4 side behind a shared NAT again. For the detail, read what CGNAT is and CGNAT and IPv6 on mobile networks.

How the IP address changes

The exit address belongs to the device's data session, not to the SIM. When the device ends its session and opens a new one, for example by toggling airplane mode or reconnecting the modem, the carrier sets up a new mapping and the device usually comes back on a different public address. This is what providers call rotation.

Two things follow. First, there is a short gap during the reconnect, so open connections drop and should be retried. Second, the new address is drawn from the carrier's own ranges: it will be a different address, but from the same network, and an address you had earlier can come back. On Coronium lines you change the IP with the dashboard button, an API call or a rotation link, or on a timer, with at least 60 seconds between changes (Brazil lines rotate on request only). More in mobile proxy IP rotation explained and what a sticky session is.

4G vs 5G mobile proxies

The radio generation changes the link between the device and the cell site. It does not change what kind of address the website sees: on 4G and 5G alike the exit is a carrier address behind the same kind of NAT.

If you need a specific network, check which carriers sell 5G lines and which are in stock today on the mobile proxy stock and price index, or read the overview of 5G mobile proxies.

Dedicated vs shared mobile proxies

Providers sell mobile proxies in two shapes. A dedicated line is one device rented to one customer: nobody else sends traffic through it, you decide when the address changes, and the bandwidth of the device is yours. A shared pool spreads many customers over many devices, so an exit address can carry other people's traffic at the same time, and rotation follows the pool rather than your workflow.

Shared pools are usually billed per gigabyte and suit light, bursty jobs. Dedicated lines are billed per period and suit long sessions, logged-in accounts and anything where another customer's behaviour on your address would hurt you. Coronium sells dedicated lines only. Each line is one physical device with its own SIM, rented to one customer. See dedicated mobile proxies for the full comparison.

Mobile vs residential vs datacenter proxies

All three are proxies. They differ in whose network the exit address belongs to, and that decides how websites treat it.

Residential pools are only as good as the way their home connections were recruited, which is worth asking any provider about; how residential proxies are sourced explains the models. For deeper comparisons read mobile vs residential proxies, ISP vs mobile proxies and what a residential proxy is.

What a mobile proxy does not change

A mobile proxy changes the network a website sees. It does not change anything your own software sends. Cookies and logins tie sessions together no matter which address they come from. Browser fingerprints (canvas, fonts, screen, WebGL, time zone) are read inside the page. TLS fingerprints come from your client library. And the TCP/IP fingerprint of the connection can reveal which operating system built the packets: a desktop OS behind a phone address is a mismatch some sites look for. Some Coronium lines offer a p0f OS fingerprint setting for this, and the stock index shows which carriers have it. Read TCP/IP fingerprinting and p0f and can mobile proxies be detected before you rely on the address alone.

How to check that a proxy is mobile

Test the exit address, not the entry address. Send one request through the proxy to an address echo, then look the result up:

curl -x http://LOGIN:PASSWORD@HOST:PORT https://www.coronium.io/api/whoami
curl -x socks5h://LOGIN:PASSWORD@HOST:PORT https://www.coronium.io/api/whoami

Paste the ip value into the IP type checker. A mobile line comes back as a mobile carrier network, or as a carrier with mobile and fixed customers where the operator mixes both under one ASN. A hosting network means you tested the gateway, or the traffic leaves through a tunnel somewhere else. If the checker shows an IPv4 carrier address, it also tells you that the address is most likely shared through carrier NAT, which is what you want to see.

Ready to try a line? Compare mobile proxies by country and carrier, check current availability on the stock index, and start with a 24-hour line on your real workflow before you commit to a month.