Quick answer
How accurate is IP geolocation at city level?
IP geolocation is reliable at country and state level and approximate at city level: databases estimate a location from routing and registration data, so two vendors can place one residential IP in a city and in its suburb. MetroProxies routes city requests from $0.50/GB (October 2026); check each exit in your target's database. No KYC; crypto.
Where does a database get a city from?
No US internet provider publishes "this IP is in this home". Geolocation vendors combine several weaker signals:
- Registration and routing: which provider holds the IP block and where its network equipment sits.
- Network measurements: response times from known points, which narrow the area.
- Provider hints: some ISPs publish geofeeds that tie IP ranges to a region or city.
- Observed data: apps and websites that share location consent data with the vendor.
Each vendor weighs these differently and updates on its own schedule, which is why two lookups of one IP can disagree.
What level of precision can I expect?
| Level | Typical agreement between databases | Use it for |
|---|---|---|
| Country | Near-total for US residential IPs | National content, currency, language |
| State | Very high | Taxes, state-level rules, statewide samples |
| Metro / DMA | High | TV and ad markets, regional prices |
| City | Good in large cities, weaker in suburbs and small towns | Local search, store pickers, city ads |
| ZIP code | Approximate; often a neighboring ZIP | Getting close to a neighborhood, then verifying |
Why do suburbs show up as the big city?
Home internet providers often serve a whole metro from shared equipment, and many databases label that equipment with the best-known city nearby. So an IP in Garland or Mesquite can read as Dallas, one in Hialeah as Miami, one in Chandler as Phoenix.
The city pages show which suburbs share a metro area: the Dallas-Fort Worth-Arlington metro alone has 11 cities with a page here. When your target labels everything "Dallas", targeting a suburb gives you a different neighborhood but the same label. When your target distinguishes suburbs, ask for the suburb.
How do I test accuracy for my own target?
- Pick 5 to 10 cities you care about.
- For each, open 5 fresh connections with the
city-segment and record the exit IP. - Look each IP up in the database your target uses (if you know it) and in one general service.
- Load the target page and note the location it shows (store, delivery area, local results).
- Count how often the target agrees with the city you asked for.
That ratio is the number that matters, and it varies by site. Rerun the test monthly; databases update constantly. The verification guide has the code for steps 2 and 3.
Where should I go next?
Test your own exits with the proxy location checker, narrow requests with proxy by ZIP code, and pick suburbs from the city list. More on the guides page.
What else do people ask about accuracy?
Can MetroProxies guarantee that every database shows the requested city?
No provider can, because the databases are run by third parties. We route the request by city; how each database labels the IP is its own estimate. Check exits in the database your target uses.
Does the time zone or language reveal a different place?
A page can read your browser's time zone and language, which come from your device, not from the IP. Set them to match the city you test (each city page gives its IANA time zone) so that the page renders the local version consistently.
Are mobile IPs located more precisely?
Usually less. Mobile carriers route many phones through a few gateways, so mobile IPs are often placed in a large city far from the phone. For city-level work, residential IPs are the better choice.
Sources (2)
- RFC 8805: A Format for Self-Published IP Geolocation Feeds — IETF (2026-10-18)
- Metropolitan and micropolitan statistical areas totals: 2020-2024 — US Census Bureau (2026-10-18)