How our IP data is built

Database last updated: August 8, 2026 · 3,535,370 curated IPv4 ranges

Where the data comes from

Every IP address record on this site is assembled from several independent sources, cross-checked against each other. No single feed is trusted on its own.

Regional internet registries

Allocation and assignment records from RIPE, ARIN, APNIC, LACNIC and AFRINIC tell us who an address block is registered to, in which country, and under which network name. We parse the full registry databases, not just per-query WHOIS.

Operator-published geofeeds

Many ISPs publish geofeed files (RFC 8805) that map their own ranges to cities. We collect hundreds of thousands of these entries, but apply them only after checking that the feed is fresh, referenced in the registry, and consistent with what the network itself shows.

Reverse DNS analysis

Hostnames assigned by operators often encode the connection type and the city of the access node. A pool named after a DSL access concentrator in a specific town is strong evidence of both technology and location. We resolve and classify millions of these hostnames.

Routing and RPKI

Live BGP data tells us which network actually announces a prefix today, which catches stale registry records. RPKI validation flags announcements that are not cryptographically authorized by the address holder.

Commercial and community data

We include GeoLite2 data created by MaxMind, available from maxmind.com, as one input among several, and community abuse reports for the security sections.

How records are verified

Sources disagree with each other constantly, and that is where most of the work is. When two of the three main signals (registry, geofeed, reverse DNS) agree and one disagrees, the outlier is corrected. Changes are applied per network, never by blanket pattern matching, and every change batch is logged and reversible. Ranges where the evidence is genuinely ambiguous are left untouched rather than guessed.

Ownership details such as the operating brand, parent company and connection type are researched manually against the operator's own published information before being recorded.

Accuracy expectations

Country-level accuracy is effectively solved for allocated space. Region and city accuracy depend on how much an operator publishes about its own network: strongest in North America and Europe, weaker where geofeeds are sparse. Mobile, satellite and VPN ranges are inherently less precise because the address is shared or intentionally relocated. City-level results should always be read as the location of the access infrastructure, not of a person.

Corrections

If you operate a network and see wrong data on your ranges, send the correct values or your geofeed URL through the contact page. Verified corrections are usually applied within a day.