A proxy server is a computer that sits between you and the internet: your request goes to the proxy, the proxy forwards it to the website, and the website's answer comes back the same way. The site you visit sees the proxy's IP address, not yours. That single property — substituting one network identity for another — is what every use of proxies is built on, from checking search results in another country to running hundreds of accounts without them being tied together.
What happens to a request, step by step
Without a proxy, opening a page is a direct conversation. Your device resolves the site's domain name to an IP address, opens a connection to that IP, sends an HTTP request, and receives the response. The site's server logs your IP, your User-Agent, and a handful of connection details.
With a proxy the sequence gains one hop:
- Your application (browser, script, mobile app) is configured with the proxy's address and port — for example
203.0.113.10:8080. - Instead of connecting to the target site, the application connects to the proxy and says, in effect, “fetch this URL for me.”
- The proxy opens its own connection to the site, using its own IP address, and relays the request.
- The site answers the proxy; the proxy passes the answer back to you.
From the site's point of view, the visitor is the proxy. From your point of view, nothing changed except one setting. Everything else about proxies — types, protocols, pricing, detection — is a variation on this one exchange.
Forward proxy vs reverse proxy
The word “proxy” is used for two opposite arrangements, and confusing them is the most common beginner mistake.
- Forward proxy works for the client. You configure it on your side; it hides you from the servers you visit. Every proxy sold for scraping, multi-accounting, ad verification or geo-testing is a forward proxy. This article — and this whole site — is about forward proxies.
- Reverse proxy works for the server. The website's owner puts it in front of their infrastructure to balance load, terminate TLS, cache pages, or filter attacks. Cloudflare and Nginx in front of a web app are reverse proxies. Visitors do not configure anything and usually cannot tell it is there.
When a provider sells you “proxies,” they are selling forward proxies. When an anti-bot vendor says a site is “behind a proxy,” they mean a reverse one. Same word, opposite direction.
The main types of proxy you will meet
Proxies are classified along three independent axes. Any real proxy is a combination of one choice from each.
By protocol: HTTP, HTTPS, SOCKS5
The protocol defines how your application talks to the proxy. An HTTP proxy understands web requests and can read, cache or modify them; an HTTPS proxy (technically an HTTP proxy using the CONNECT method) tunnels encrypted traffic without seeing inside it; a SOCKS5 proxy works one level lower and forwards any TCP or UDP traffic — web, email, games, torrents — without understanding it. If you only ever use a browser, HTTP(S) is enough. If you need to proxy an arbitrary application, you want SOCKS5. The full trade-off is in our SOCKS5 vs HTTP comparison.
By IP origin: datacenter, residential, mobile
This is the axis that determines price and how “trustworthy” the IP looks to the site you visit.
- Datacenter IPs belong to hosting companies. Fast, cheap, plentiful — and easy to recognise, because the whole IP range is registered to a cloud provider, not to an internet subscriber.
- Residential IPs are assigned by consumer ISPs to real households. Traffic from them looks like an ordinary person at home. Slower and pricier, usually billed per gigabyte.
- Mobile IPs come from cellular carriers. Thousands of real phones share each IP through carrier-grade NAT, so sites cannot block one without blocking many legitimate users. Highest trust, highest price.
We compare all three, with a table, in Residential vs datacenter vs mobile proxies.
By behaviour: static or rotating, shared or dedicated
A static proxy gives you the same IP for as long as you rent it; a rotating one changes IP on a schedule or on every request. A dedicated proxy is yours alone; a shared one is used by several customers at once, which is cheaper but means you inherit whatever reputation the others earned. Rotation is its own topic — see What is IP rotation.
What a proxy hides — and what it does not
A proxy replaces your IP address. It does not automatically make you anonymous, and assuming it does is how most bans happen.
| Signal | Hidden by a proxy? | Notes |
|---|---|---|
| Your IP address | Yes | The site sees the proxy's IP |
| Your approximate location | Yes | Geolocation follows the proxy's IP |
| Contents of HTTPS traffic | Not applicable | Encrypted end-to-end; the proxy relays it blind (unless you install its certificate) |
| Cookies and logins | No | Sent by your browser regardless of IP |
| Browser fingerprint (fonts, canvas, WebGL, screen) | No | Unchanged; sites correlate it across IPs |
| TLS fingerprint (JA3/JA4) | No | Depends on your client software, not the IP |
| WebRTC local IP leak | No | Browsers can reveal your real IP unless WebRTC is disabled |
| DNS requests | Depends | Only if DNS is resolved through the proxy (SOCKS5 with remote DNS, or the proxy's own resolver) |
In other words, a proxy answers the question “where is this request coming from?” It says nothing about “who is sending it” or “is this the same person as before?” Sites answer those from cookies, fingerprints and behaviour. That is why serious multi-account work pairs proxies with an antidetect browser, and why you should test a proxy for leaks before trusting it.
Proxy vs VPN in one paragraph
A VPN does the same IP substitution but for all traffic from your device, at the operating-system level, inside an encrypted tunnel. A proxy is configured per application and usually per task. VPNs are for privacy of a person; proxies are for control of traffic — many IPs, many sessions, chosen per request. The longer answer, including Tor, is in Proxy vs VPN vs Tor.
When you actually need a proxy
- Seeing the internet from somewhere else — checking prices, ads or search results as a user in another city or country would see them.
- Collecting data at scale — a scraper making thousands of requests from one IP is throttled or blocked within minutes; spreading them across many IPs keeps each one below the radar.
- Running several accounts on one platform — social networks, marketplaces and ad platforms link accounts that share an IP. Each account on its own IP is the baseline for not being linked.
- Testing your own service — QA from multiple regions, verifying geo-blocking works, load-testing from distributed sources.
- Bypassing an IP-based block on a network you use, where the block is a nuisance rather than a policy you are bound by.
If your need is simply “I do not want my ISP or a café Wi-Fi to see what I browse,” a VPN is the simpler tool. Proxies earn their keep when you need many identities, chosen per task, from specific places.
How to get started
- Decide what you are proxying: a browser only (HTTP is fine) or any application (SOCKS5).
- Decide how trustworthy the IP must look: datacenter for speed and volume where blocks are rare, residential or mobile where the target site is aggressive about bots.
- Buy access — you receive a host, a port, and either a username/password or an IP whitelist.
- Configure the application. Our step-by-step setup guide covers Chrome, Firefox, Windows and macOS.
- Verify it: open an IP-check page and confirm the IP, location and DNS resolver have all changed, and that WebRTC is not leaking your real address.
For a shortlist of providers by use case, see our ranking of the most reliable proxy providers.
Frequently asked questions
Is using a proxy legal?
Using a proxy is legal in almost every country; it is a standard networking tool used by corporations, schools and ISPs. What may be prohibited is what you do through it — violating a platform's terms of service, evading a legal geo-restriction, or unauthorised access. The proxy is not the offence; the activity might be.
Does a proxy make me anonymous?
No. It hides your IP address from the sites you visit. It does not hide your cookies, logins, browser fingerprint or behaviour, and the proxy operator itself can see your traffic (except the encrypted contents of HTTPS). Anonymity is a combination of tools and habits; a proxy is one component.
Is a proxy the same as a VPN?
Both replace your visible IP. A VPN encrypts and routes all device traffic through one tunnel and is set up once at the OS level; a proxy is configured per application and lets you use many different IPs for different tasks. VPNs suit personal privacy, proxies suit controlled multi-IP work.
Will a proxy slow down my connection?
Every hop adds latency, typically 20–200 ms depending on distance and load. Datacenter proxies are fastest; residential and mobile are slower because the traffic really does travel through a home or a phone. For browsing the difference is rarely noticeable; for high-volume scraping it is a real throughput factor.
Can a website tell I am using a proxy?
Often, yes. Datacenter IP ranges are publicly known and easy to flag. Residential and mobile IPs are much harder to distinguish from ordinary visitors by IP alone, which is why they cost more. Beyond the IP, sites look at fingerprints and behaviour, which a proxy does not change.
Free proxies — are they safe?
Generally not. Free public proxies are slow, short-lived, shared with thousands of strangers (so their IPs are already blacklisted), and some are run specifically to inspect or inject into your traffic. For anything involving logins or money, use a paid provider with a stated privacy policy.