Language
Log in Sign up

How to Test a Mobile Proxy (Step by Step)

Testing a mobile proxy means confirming five things in under five minutes: the exit IP changes when you route through it, that IP belongs to a mobile-carrier ASN (not a datacenter), the geolocation matches the country you ordered, latency is usable, and rotation actually issues a new IP. A real 4G/LTE/5G proxy returns a carrier IP behind CGNAT with a 100% trust score — a fake one returns a hosting-provider IP.

A mobile proxy passes the test when an IP-info lookup made through the proxy returns an IP whose ASN belongs to a mobile network operator (for example AT&T, T-Mobile, Verizon in the US, or Free Mobile, SFR and Orange in France), flags the connection as mobile/cellular, and resolves to the country you paid for. Proxy4G runs every IP on a physical SIM across 43 carriers in 18 countries, so a correct test shows a carrier ASN and a CGNAT exit — never a hosting provider like OVH, Hetzner or DigitalOcean.

What does a passing test actually prove?

Five independent signals separate a genuine mobile proxy from a relabelled datacenter IP. Run all five — passing one or two is not enough, because some sellers proxy datacenter traffic through a residential-looking front and still leak on ASN.

  • Exit IP changes: the IP seen by a target site is the proxy's, not your own.
  • Carrier ASN: the autonomous-system owner is a mobile network operator, and the connection type is reported as mobile/cellular.
  • Geolocation: the IP resolves to the country (and roughly the region) you ordered.
  • Latency: round-trip time is consistent and workable for your task — mobile adds some delay versus datacenter, which is expected and normal.
  • Rotation: triggering a rotate (on demand or on a timer) yields a different exit IP within the same carrier/country.

If the ASN says "hosting" or names a server company, stop — it is not a mobile proxy regardless of marketing. See mobile vs residential vs datacenter for why the ASN is the decisive signal.

How do you test a mobile proxy step by step?

  1. 1

    Connect with your credentials

    Use the host, port, username and password emailed within minutes of payment, in the form HOST:PORT. Pick HTTP, HTTPS or SOCKS5 — all plans support all three, with username/password or IP-whitelist auth.

  2. 2

    Check the exit IP

    Send a request through the proxy to an IP-echo endpoint such as https://api.ipify.org. Confirm the IP returned is NOT your own home/office IP. This proves traffic is actually leaving via the proxy.

  3. 3

    Confirm the IP type and ASN

    Look up that IP with an IP-info service and read the ASN/organisation and connection type. It must name a mobile carrier and report 'mobile' or 'cellular' — not 'hosting', 'datacenter', or a server-provider name. This is the single most important check.

  4. 4

    Verify geolocation

    Confirm the country (and broad region) matches what you ordered — e.g. a France plan resolves to FR on Free Mobile, SFR, Orange or Bouygues. Note that CGNAT means city-level geo can be approximate; country is what counts.

  5. 5

    Measure latency

    Time a few requests through the proxy (e.g. curl -w time_total). Expect mobile-network overhead versus datacenter; look for stable, repeatable timings rather than the lowest possible number.

  6. 6

    Test rotation

    On a Dedicated plan, trigger rotate-on-demand (dashboard or reset link) or wait for your 1–60 min timer; on Shared, the IP auto-rotates every 5 minutes. Re-run the IP check and confirm you get a new carrier IP in the same country.

What commands test a mobile proxy with curl?

bash
# 1) Exit IP through the proxy (replace placeholders with your emailed creds)
curl -x http://USER:PASS@HOST:PORT https://api.ipify.org

# 2) Full IP intel: ASN, org and connection type
curl -x http://USER:PASS@HOST:PORT https://ipinfo.io/json

# 3) SOCKS5 instead of HTTP (same credentials)
curl -x socks5h://USER:PASS@HOST:PORT https://ipinfo.io/json

# 4) Measure latency through the proxy
curl -x http://USER:PASS@HOST:PORT -o /dev/null -s -w "connect=%{time_connect}s total=%{time_total}s\n" https://api.ipify.org

# 5) Test rotation: capture IP, rotate, capture again
curl -s -x http://USER:PASS@HOST:PORT https://api.ipify.org; echo
# (trigger rotate via dashboard/reset link, or wait for the timer)
curl -s -x http://USER:PASS@HOST:PORT https://api.ipify.org; echo

What does a real carrier ASN look like?

An ASN (Autonomous System Number) identifies who owns a block of IPs. For a genuine mobile proxy the ASN's organisation is a mobile network operator, and reputable IP-intel databases tag the connection as mobile or cellular. Examples you might see by country on Proxy4G:

  • United States: AT&T, T-Mobile US, Verizon Wireless (plus T-Mobile 5G).
  • United Kingdom: Three (Hutchison 3G), EE.
  • France: Free Mobile, SFR, Orange, Bouygues Telecom.
  • Germany / Spain / India: O2 and Vodafone (DE); Movistar, Orange, DIGI (ES); Vodafone and Jio (IN).

By contrast, a datacenter IP shows an ASN owned by a hosting company (OVH, Hetzner, DigitalOcean, AWS, Google Cloud) and is tagged hosting/datacenter. That distinction — carrier vs hosting — is the verdict. Definitions for ASN, CGNAT, trust score and more are in the glossary.

Test-passed checklist

  • Exit IP through the proxy differs from your own public IP
  • ASN/organisation names a mobile carrier, not a hosting provider
  • Connection type reads 'mobile' or 'cellular'
  • IP geolocates to the country you ordered
  • Latency is stable and repeatable across several requests
  • Rotation (on-demand, timer, or the 5-min Shared cycle) returns a fresh carrier IP
  • No DNS or WebRTC leak exposes your real IP (use socks5h:// to push DNS through the proxy)

Why mobile proxies behave differently under test

Mobile networks place thousands of subscribers behind a handful of public IPs using Carrier-grade NAT (the RFC 6598 100.64.0.0/10 shared space — see RFC 6598). That is exactly why mobile IPs earn a high trust score: blocking one would block many real humans, so platforms treat them gently. Practical consequences for testing: city-level geolocation can look fuzzy (country is reliable), latency is naturally higher than datacenter, and the same IP may already be shared with legitimate phone users. None of these are faults — they are the fingerprint of a real 4G/5G connection. If you need protocol details for your client, SOCKS5 with socks5h:// is the safest choice because it routes DNS through the proxy too.

Common false negatives (and how to avoid them)

A test can "fail" for reasons that have nothing to do with the proxy:

  • DNS leak: with SOCKS5, use socks5h:// (not socks5://) so name resolution goes through the proxy, otherwise your local resolver can expose your country.
  • Stale ASN database: different IP-intel providers update at different times; check two sources before concluding an IP is datacenter.
  • IP-whitelist vs user/pass: if you whitelisted an IP and your machine's IP changed, auth will fail — switch to username/password to test quickly.
  • Rotation timing: on a Dedicated timer set to, say, 30 minutes, the IP won't change until the interval elapses — use the on-demand reset link to force it during testing.
  • Caching: some IP-info endpoints cache; add a cache-buster query string or test a couple of endpoints.

Once all five signals pass, you have verified a real carrier-grade mobile proxy. See how to choose a provider to turn these checks into buying criteria.

Frequently Asked Questions

What is the single most important check when testing a mobile proxy?

The ASN. Route an IP-info lookup through the proxy and read the autonomous-system owner and connection type. A real mobile proxy names a carrier (AT&T, T-Mobile, Free Mobile, Orange, Vodafone, Jio, etc.) and reports the connection as mobile or cellular. If the ASN names a hosting company (OVH, Hetzner, DigitalOcean, AWS) or reads datacenter, it is not a mobile proxy regardless of how it is marketed.

How do I test proxy rotation?

Capture the exit IP, trigger a rotation, then capture it again. On a Proxy4G Dedicated plan you rotate on demand via the dashboard or a reset link, or set an automatic timer of 1–60 minutes. On a Shared plan the IP auto-rotates every 5 minutes. A correct rotation returns a different IP that still belongs to the same carrier and country.

Why is my mobile proxy slower than a datacenter proxy?

Because it routes over a real cellular network behind carrier-grade NAT, which adds latency a datacenter never has. That overhead is the trade-off for a genuine carrier IP with a 100% trust score. When testing, look for stable, repeatable timings across several requests rather than the absolute lowest number — consistency matters more than raw speed.

Why does the geolocation look slightly off?

CGNAT places many subscribers behind shared carrier IPs, so city- or region-level geolocation is often approximate even though the country is correct. Treat country as the reliable signal. If your test shows the right country on the carrier you ordered, the proxy passes — exact-city accuracy is not something any real mobile proxy can guarantee.

How do I avoid a DNS leak when testing with SOCKS5?

Use the socks5h:// scheme rather than socks5://. The extra h tells curl (and most clients) to resolve hostnames through the proxy instead of your local resolver, so DNS queries don't expose your real location. All Proxy4G plans support SOCKS5 alongside HTTP and HTTPS, with username/password or IP-whitelist authentication.

Can I test the proxy before paying?

Credentials (host, port, username, password) are emailed within minutes of payment, and accounts are auto-created on your first order with no KYC. Pay with Bitcoin, Ethereum, Solana or USDT, then run the curl checks immediately. Start small with a 1-month Shared plan from $10.80 to validate carrier ASN, geo and rotation before committing to a longer Dedicated configuration.

Test a real carrier IP in minutes

Order a Shared plan from $10.80/mo or Dedicated from $27/mo, pay with crypto, and run the curl checks the moment your credentials land. No KYC.

Build your proxy