WhatWebSees Learn
How IP Geolocation Works
Published and updated 2026-09-21
IP geolocation estimates where an Internet address is used. It is a database lookup, not a GPS measurement. A provider combines network registration, routing, infrastructure observations and correction data into a mapping from address ranges to places. When a tool reports a city, it is returning the database’s current best label for that range—not locating a device in real time.
From network allocation to a location estimate
Address blocks are allocated through regional Internet registries and then assigned or delegated to networks. Public registration records can reveal an organization and broad service region, but their postal contact is not necessarily where users are. Geolocation vendors refine that foundation with routing patterns, ISP information, network topology, latency observations, user-submitted corrections and other licensed or public sources. The exact inputs and update schedules differ by database.
Mappings normally apply to a prefix rather than one uniquely researched address. If an ISP uses a block across several cities, the database may choose a central or operational location. A “city” can therefore mean likely service area, gateway location, corporate office, or a default point chosen because nothing more precise is justified.
Typical confidence by level
| Field | How to interpret it |
|---|---|
| Country | Often the strongest geographic estimate, but still affected by roaming, VPNs and cross-border networks. |
| Region | Useful when allocation and network topology are regional; not guaranteed. |
| City | An approximate label with potentially wide error, especially for mobile and centralized providers. |
| Coordinates | Usually a representative point for the database’s city or region, not device coordinates. |
| Time zone | Derived from the estimated place and unreliable near boundaries or for remote exits. |
The WhatWebSees IP lookup uses a local database rather than sending the submitted address to an external geolocation API. Its deterministic fields—address version and routing classifications—are different from its approximate location fields.
Why databases disagree
There is no single authoritative table assigning every public address to a city. Providers ingest different evidence, group prefixes at different sizes, resolve conflicting signals differently and publish on different schedules. An ISP can move a block before every database catches up. Small prefixes can also inherit the location of a larger parent block when finer evidence is unavailable.
Corrections have tradeoffs. A network operator knows its allocation, while an end user knows where a connection is used; both observations can become stale. Fraud can make self-reported evidence unreliable. Conservative databases may deliberately prefer a less precise location rather than assert a city from weak evidence.
VPNs, proxies and hosting networks
A destination sees the address that connects to it. With a VPN, that is usually the VPN exit; with an explicit proxy, it is the proxy; with a corporate secure gateway, it may be a company egress point. A residential proxy may look like an ordinary ISP connection. Tor traffic appears from an exit relay. Geolocation can describe the visible network endpoint reasonably while saying little about the person behind it.
Cloud services create another distinction. An address might be registered to a global provider, routed with anycast, or serve infrastructure in multiple regions. Registration country, route origin, physical server and current user location can all be different concepts.
Mobile and satellite networks
Mobile carriers often centralize traffic through packet gateways. A phone in one region may emerge onto the public Internet in another, and roaming can send traffic through a home or partner network. Carrier-grade NAT makes many subscribers share an address. Rapid lease turnover also weakens long-lived assumptions.
Satellite systems can route a terminal through a distant ground station or provider gateway. The public address may map to that infrastructure even though the dish is elsewhere. Similar effects occur on aircraft, ships and some fixed wireless services.
Why IP geolocation is not GPS
GPS and device location services estimate the position of hardware using satellite signals, nearby Wi-Fi, cellular data and sensors—usually only after an application receives permission. An IP lookup sends no such coordinates. Latitude and longitude in a GeoIP result are commonly the center of a city or another representative point selected by the database.
That difference has practical consequences. GeoIP may be suitable for setting a default language, selecting a broad content region, troubleshooting traffic routing or highlighting an unusual country. It is unsuitable for emergency location, property-level decisions, proof of residence, or claims that a person was at a precise place.
Using the result responsibly
Label the location as approximate, preserve the raw country/region distinction, and expect unknown values. If a decision has meaningful consequences, allow correction and use corroborating evidence. Do not silently convert a city estimate into a street-level claim.
Results can change when a database updates, an ISP reallocates a prefix, the user changes networks, or a privacy service changes the visible exit. That variability is normal. For the addressing background behind these mappings, read What Is an IP Address?.