Your public IP address

How this page finds your IP address

Every connection to a website starts with a packet that carries a return address. That return address is your public IP address, and the server on the other end must know it to send the page back. This site simply reads the address your request arrived from and prints it. No script has to run, no permission is asked for, and no third-party service is involved in that first step.

What you see is the address of the last device that spoke to us on your behalf. For a home connection that is your router's public side, assigned by your ISP. On a mobile phone it is an address shared by many subscribers behind the carrier's gateway. If you are on a VPN, a proxy or the Tor network, it is the exit point of that service, not your own line. That is the whole point of those tools: the number here is the number every other website sees too.

The page also tries to learn a second address. Most browsers prefer IPv4 when both protocols work, so the server-side address is usually IPv4. A small script then asks our IPv6-only endpoint for the address you use on that protocol. If the request succeeds, the IPv6 line in the header fills in. If your network has no IPv6, it says so. Both addresses matter for privacy, because a tunnel that protects one and ignores the other leaves a gap. The IPv6 test explains that case in detail.

How your location, ISP and network type are found

Nothing in an IP address encodes a place. Location comes from a database that maps address blocks to the region where the holder registered them. We use the DB-IP Lite City database, refreshed every month. It gives us a country, a region, a city estimate, coordinates and a time zone. Country is reliable because address blocks are handed out by regional registries and ISPs rarely use a block outside their own country. City is weaker: it reflects where the block was registered or where most of its traffic appears, which may be the ISP's head office rather than your street.

The ISP and the network come from public routing data. Every address on the internet is announced by an autonomous system, a network with its own number, and that announcement is public. We import the global table of announced prefixes every week and match your address to the block that contains it. The result is the ASN, its registered name and the prefix it belongs to, for example a /22 that the ISP announces for a whole city.

The network type is a classification layered on top of that. A consumer ISP, a mobile carrier, a hosting company and a university all announce addresses in the same way, so we read the network's name and compare it with keyword lists for hosting, mobile, education and government networks, then check the address against a published list of datacenter ranges, the daily Tor exit list and the Spamhaus DROP list. The label you see, such as "Residential", "Mobile carrier" or "Datacenter / hosting", is the outcome of those checks. It is a heuristic, and we say so on every page, but it is the same heuristic most fraud filters apply in a stricter form.

How to read your result

The pill above the address summarizes the 12 signals we score on every visit. The signals that depend only on the network are known as soon as the page loads; the ones that depend on your browser fill in once the page scripts run. The full list with point values lives on the anonymity score page. Here is what the three verdicts mean.

The classification next to your ISP name is a separate, simpler label that describes only the network:

If you are checking a proxy or VPN before logging into an account, the two rows that matter most are the network type and the time zone comparison. A residential IP with a browser clock that matches its country looks like a normal visitor. A hosting address, or a residential address with a clock three time zones away, is the pattern that gets a login challenged.

How to fix what is flagged

Each marked row on this page links to a short explanation of the signal and its fix, and the dedicated tests are linked below. In the order that most often resolves a bad verdict:

1. Datacenter or hosting network. No browser setting changes this; the address itself is the problem. Switch to a residential or mobile proxy or to a VPN server that exits on a consumer network, then reload this page and confirm the label changed.
2. Time zone vs IP location. Set your operating system clock to the time zone of the IP's country. On Windows, open Settings, then Time and language, then Date and time, turn off automatic time zone and pick the zone. On macOS, open System Settings, then General, then Date and Time. On Android and iOS, turn off automatic time zone under Date and time settings. In an antidetect browser, set the profile's time zone to match the proxy instead of changing the system clock.
3. WebRTC public IP. If WebRTC exposes a different address, the browser is talking to the internet outside your tunnel. Run the WebRTC leak test for per-browser steps; most VPN clients and antidetect browsers have a switch for it.
4. DNS resolver location. If your DNS queries are answered from another country, set the resolver in your VPN client or operating system to one the tunnel carries. The DNS leak test shows the resolvers in use.
5. IPv6 address. If an IPv6 address from a different country is reachable while your IPv4 goes through a tunnel, disable IPv6 on the adapter or enable IPv6 in the VPN client. See the IPv6 test.
6. Proxy headers. Headers such as Via or X-Forwarded-For are added by the proxy software, not by you. Switch to a proxy that runs in elite or anonymous mode, or ask the provider; the HTTP headers page shows exactly what was received.
7. Browser language vs country. A weak signal on its own, but easy to align: add the local language of the IP's country to your browser languages or leave English first, which always passes.
8. Automation signals. A headless or webdriver-controlled browser is marked as automation. Use a normal browser profile for logins.

Limits of this check

• City is an estimate. The geolocation database tells us where an address block was registered or where its traffic clusters, not where you are sitting. Expect the city to be wrong some of the time and the country to be right almost always. The database is refreshed monthly, so a block that moved last week may still show its old place.
• Network type is a heuristic. We classify from the network's name, a datacenter list, the Tor exit list and the Spamhaus DROP list. A hosting company with an innocent name can slip through as "Unclassified", and a small ISP whose name contains "cloud" can be mislabeled. The label tells you how an automated filter is likely to read the address, not what the address is used for.
• We only see what reaches us. If the site you care about sits behind a different edge network, uses a commercial fraud database or keeps its own history of your account, it may reach a different conclusion with the same address.
• Browser signals need JavaScript. Time zone, WebRTC, IPv6 and DNS results only appear when the page scripts run. Without them the verdict is based on the network and request headers alone and the page says how many signals were checked.
• A clean verdict is not a guarantee. It means none of the 12 signals we check raised a flag. It does not mean a website cannot recognize you by other means, such as cookies, an existing account history or a payment instrument.

Why does my IP show the wrong city?

This is the question people ask most, so it deserves a full answer. An ISP receives large address blocks from its regional registry and splits them across its own network as it sees fit. The registration record often names the company's head office or the city where the block was first deployed. When the ISP later moves a subnet to another town, nothing forces it to tell anyone. Geolocation providers compensate by watching where traffic from each block appears, which works well for large cities and poorly for everything else.

Mobile networks add another layer. Phones share a pool of public addresses behind carrier-grade NAT, and that pool may sit in a regional hub hundreds of kilometers from the user. A subscriber in one city routinely appears in the carrier's gateway city. VPNs and proxies are a third case: the address is wherever the exit server is, which is exactly what you want, but the database may still name a different city than the provider advertises, because the provider's marketing and the registry record disagree.

If the city is wrong for your home connection, there is not much to fix and usually no reason to try. Websites use the country far more than the city. If the city is wrong for a proxy you are paying for, compare it with the IP lookup result for the same address, which shows the registry record and the announcing network; that tells you whether the provider's claim or the database is closer to the truth.

IPv4 and IPv6: which one are you seeing?

An IPv4 address looks like four numbers separated by dots, for example 203.0.113.7. An IPv6 address is longer, written in hexadecimal groups separated by colons. Most networks still prefer IPv4 for the first connection, so the address at the top of this page is usually IPv4, and the IPv6 line is filled in by a separate probe. Having both is normal and not a problem in itself. The problem case is a tunnel that only carries IPv4 while the operating system still sends IPv6 traffic directly; a site that supports both protocols then sees your real network on one of them.

Public and private addresses

The address here is your public one. Inside your home, every device has a private address that your router hands out, typically in the 192.168.x.x range. Those private addresses never leave your network; the router translates them to the single public address when traffic goes out. If you look up an address on this site and it is private or reserved, the page tells you so and shows no location, because there is none to show.

Does my IP address change?

For most home connections, yes, but not often. ISPs assign addresses from a pool with a lease, and a reconnect can give you a new one, though many routers keep the same address for months. Mobile carriers change addresses frequently, sometimes every time the phone drops and rejoins the network. Servers, business lines and anything you pay a static-IP fee for stay fixed. If you need a different address right now, the fastest options are to switch between Wi-Fi and mobile data, restart the router and wait a few minutes, or use a VPN or proxy. Check back here afterwards; the pill at the top tells you whether the new address looks like a home connection.

What a website can and cannot learn from your IP

From the address alone, a site learns the country, usually the region, sometimes the city, the ISP or hosting company, and whether that network is one that normal visitors use. It can also check the address against blocklists and against its own records of past abuse. It cannot learn your name, your street address or your identity; only your ISP can connect the address to a subscriber, and it does so only on a legal request.

What makes the difference in practice is not the address by itself but whether the rest of your connection agrees with it. A browser that reports a time zone from another continent, a DNS resolver in a third country and a WebRTC address that differs from the public one is a pattern that fraud filters are built to catch. Each of those is a test on this site, and the IP leak test runs the connection-level ones together.

Checking a proxy or VPN before you log in

If you use a proxy, a VPN or an antidetect browser profile for an account, do the check in this order, and do it from the same profile you will log in with. First look at the network label at the top of this page: it should say residential or mobile carrier, not datacenter. Then open the anonymity score and let the browser signals load; every row that is not a pass links to its fix. Then run the DNS leak and WebRTC tests, because those are the two leaks a login page can detect without your noticing. When every row passes, reload once to confirm the result is stable, and only then open the site you care about.

Looking up someone else's address

The box labeled "Look up an IP" accepts any IPv4 or IPv6 address or a hostname. The result page shows the same location, network and classification data we show for your own address, plus the registry record and reverse DNS name. It is useful for checking a proxy before you buy it, reading a server log, or confirming the address in an email header. Lookups are rate limited to 30 per minute per visitor, and the same data is available as JSON from the public API for scripts.