Guide · Proxy basics
How to test a proxy's location
To test a proxy's location, send a request through it to an IP-echo service to get the exit IP, then look that IP up in one or more geolocation databases — with curl that is a single command. Expect the country to agree almost always and the city to disagree often: geolocation is an estimate, and one major vendor puts its own US city-level accuracy at about 66% within 50 km. The test that matters most is what the target site shows: its language, currency and regional content.
Reviewed 30 Oct 2026
The short answer
- Exit IP first:
curl -x http://USER:PASS@GATEWAY https://api.ipify.orgreturns the address the internet sees. - Country-level geolocation is reliable — MaxMind estimates 99.8% — but US city-level accuracy is about 66% within 50 km in its own estimate.
- Databases disagree because they are estimates built from different data, updated at different times; mobile and carrier-NAT addresses are coarser still.
- Network operators can publish where their prefixes are (RFC 8805 geofeeds), but databases adopt changes on their own schedules.
- Verify against the target itself: the market, language and currency it serves are what your data depends on.
Step 1: confirm the exit IP
Send a request through the proxy to a service that echoes your address. Do it twice with a rotating username — you should usually see two addresses — and twice with a sticky one, where you should see the same address both times. That one test confirms credentials, targeting and session behaviour together.
Step 2: look up the geolocation
Look the exit IP up in a geolocation service. Most return country, region, city, the network's organisation and its autonomous system number. Some, such as MaxMind's, also return a connection type and a user type — residential, cellular, hosting and so on — which is a quick way to confirm you got the kind of proxy you paid for.
| Check | How | What it tells you |
|---|---|---|
| Exit IP | Request an IP-echo URL through the proxy | Credentials work; which address the internet sees |
| Country | Geolocation lookup on that IP | Whether country targeting worked — should agree across databases |
| City / region | Same lookup; compare two databases | An estimate; disagreement is normal |
| Network type | ASN, organisation, user or connection type | Residential vs hosting vs cellular |
| Target's view | Load the real page; read language, currency, region | What your collected data will actually reflect |
Why geolocation databases disagree
IP geolocation is inference, not a lookup in an official register. Vendors combine registry data, network measurements, operator feeds and user-submitted corrections, and each weighs them differently. MaxMind, one of the most widely used, estimates its own accuracy at 99.8% for countries, about 80% for US states and about 66% for US cities within a 50 km radius — and returns an accuracy radius with each city-level answer.
Some addresses cannot be placed precisely at all. MaxMind notes that mobile network addresses may be used by phones across a large distance, so it may return only a country and state. Carriers also share addresses among many subscribers through carrier-grade NAT, a practice with its own reserved address range (RFC 6598) — see CGNAT. For mobile proxies, a city-level lookup result is not meaningful; we do not offer city targeting there.
Operators can publish self-reported locations for their prefixes as geofeeds in a simple CSV format (RFC 8805). Databases ingest them on their own schedules, so a recently moved range can show an old location in one database and a new one in another.
Test what the target sees
- Load the actual page you will collect through the same proxy string and record the language, currency, country selector and any regional notice.
- Compare with a known baseline from inside that country if you have one.
- For city-specific content — local search results, store availability — check a city name or postcode on the page, not only the IP lookup.
- If the target shows the wrong market, check the site's own country setting and cookies before blaming the IP; many sites remember a previous choice.
City targeting on our rotating pools uses the _city- token from a published list, for example USERNAME_pool-arcfast_country-de_city-berlin. How it applies per pool is on the residential proxies page and in geo-targeting.
Code
Worked example
# Exit IP through a rotating German exit (run twice: usually two different IPs)curl -s -x "http://USERNAME_pool-arcfast_country-de:PASSWORD@proxy.arcproxies.io:8000" https://api.ipify.org; echocurl -s -x "http://USERNAME_pool-arcfast_country-de:PASSWORD@proxy.arcproxies.io:8000" https://api.ipify.org; echo# Sticky session (run twice: the same IP while the session holds)curl -s -x "http://USERNAME_pool-arcfast_country-de_session-k3f9a1x2_sesstime-10:PASSWORD@proxy.arcproxies.io:8000" https://api.ipify.org; echocurl -s -x "http://USERNAME_pool-arcfast_country-de_session-k3f9a1x2_sesstime-10:PASSWORD@proxy.arcproxies.io:8000" https://api.ipify.org; echo# Geolocation as seen by one database (country, region, city, org/ASN)curl -s -x "http://USERNAME_pool-arcfast_country-de:PASSWORD@proxy.arcproxies.io:8000" https://ipinfo.io/json# Same over SOCKS5, resolving DNS through the proxycurl -s -x "socks5h://USERNAME_pool-arcfast_country-de:PASSWORD@proxy.arcproxies.io:8000" https://api.ipify.org; echoFAQ
Questions
Something not covered? Write to contact@arcproxies.io.
How do I check my proxy's location?
Send a request through the proxy to an IP-echo service to get the exit IP, then look that IP up in a geolocation database. With curl: curl -x http://USER:PASS@HOST:PORT https://ipinfo.io/json returns the IP with its country, region, city and network in one step.
Why does my proxy show the wrong city?
Because IP geolocation is an estimate. MaxMind, for example, estimates its US city-level accuracy at about 66% within 50 km, and different databases use different data and update at different times. The country is far more reliable — MaxMind estimates 99.8%.
Why do IP location websites show different results for the same IP?
Each geolocation vendor builds its own estimate from registry data, network measurements, operator geofeeds and corrections, and refreshes it on its own schedule. Country usually agrees; region and city often do not.
How accurate is IP geolocation?
Very accurate for countries and much less so for cities. MaxMind estimates 99.8% at country level, about 80% at US state level and about 66% for US cities within 50 km. Mobile network addresses are coarser still.
How can I test whether a proxy is sticky or rotating?
Make two requests to an IP-echo service with the same proxy credentials. With a rotating username you will usually see two different addresses; with a sticky session handle you should see the same address both times while the session holds.
Can I check proxy location without curl?
Yes. Any HTTP client works — set the proxy URL and request an IP-echo or geolocation endpoint. In Python requests, pass the proxy in the proxies argument and request https://ipinfo.io/json.
Sources & review
Last reviewed by the ARC Proxies team. Prices are read from the live catalog at build time.
Next
Where to go from here
Guides