Documentation Python quickstart Blog Free tools Enterprise solutions hello@quanticdata.ioLog in

IP checkers compared: ipinfo, IP2Location, ip-api and Scamalytics

Four public addresses read on four IP checkers on 20 November 2026: ipinfo, IP2Location, ip-api and Scamalytics. The country was the same in 16 of 16 answers, the city of one address came back as four different cities, and the fraud scores were 16 for a home provider address and 0 for a datacenter address.
Four public addresses read on four IP checkers on 20 November 2026: ipinfo, IP2Location, ip-api and Scamalytics. The country was the same in 16 of 16 answers, the city of one address came back as four different cities, and the fraud scores were 16 for a home provider address and 0 for a datacenter address.

IP checkers answer two different questions. ipinfo, IP2Location and ip-api are location and network databases: they say where an address is and who operates it. Scamalytics and IPQualityScore (IPQS) are risk scorers: they estimate how much of the traffic from an address looks fraudulent. One address can read differently on each.

To show what that means in practice, on 20 November 2026 we looked up the same four public addresses on four checkers and compared every field. The addresses are DNS resolvers that their operators publish: one from Comcast, one from Deutsche Telekom, Google's 8.8.8.8 and one from Hetzner. Two sit on home provider networks, two in datacenters.

Two kinds of checker

CheckerKindWhat it returns
ipinfoLocation and networkCountry, region, city, the network with its ASN, the reverse DNS name
IP2LocationLocation and networkCountry, region, city, the network, a proxy flag
ip-apiLocation and networkCountry, region, city, the network, and three flags: mobile, proxy, hosting
ScamalyticsRisk scoreA fraud score from 0 to 100, a datacenter field, proxy fields and five blacklists
IPQSRisk scoreA fraud score with connection and abuse signals; request settings such as strictness can affect the result (as our product pages describe it; not part of the run below)

The first group describes the address. The second group judges traffic, which is why two scorers can disagree: each builds its number from its own traffic, its own lists and its own model. Other checkers that customers send us, ping0 among them, are read the same way: first decide which kind of answer each field is.

What we measured: four addresses on four checkers

AddressipinfoIP2Locationip-apiScamalytics
Comcast resolver, 75.75.75.75US, Mount Laurel, AS7922 ComcastUS, Cambridge, ComcastUS, Canton, AS7922 Comcast; hosting falseUS, Northlake; score 16; datacenter No
Deutsche Telekom resolver, 194.25.0.60DE, Hannover, AS3320 Deutsche TelekomDE, Hanover, Deutsche TelekomDE, Hanover, AS3320; hosting falseDE, Hanover; score 2; datacenter Unknown
Google, 8.8.8.8US, Mountain View, AS15169 GoogleUS, Mountain View, GoogleUS, Ashburn, AS15169; hosting trueUS, Mountain View; score 0; datacenter Yes
Hetzner resolver, 185.12.64.1DE, GunzenhausenDE, Gunzenhausen, HetznerDE, Gunzenhausen; hosting trueDE, Gunzenhausen; score 5; datacenter Yes

Each lookup was made from our own connection, with the address in the URL, on ipinfo.io, IP2Location.io, ip-api.com and scamalytics.com. Three things stand out.

The country and the network agree

The country was the same in 16 of 16 answers, and so was the operator's name wherever a checker returned one. These two fields come from the address registries and from the routing table, which every database reads. When you check a proxy, they are the fields to read first: the country tells you whether the targeting worked, and the network tells you what kind of address it is.

The city is an estimate

For the Comcast resolver the four checkers named four cities: Mount Laurel, Cambridge, Canton and Northlake. That address is announced from many places at once, which ipinfo marks with anycast: true, so there is no single right city. For the Deutsche Telekom address all four said Hanover. City data is each database's own estimate; our proxy location checker shows seven databases next to each other for that reason.

For the proxies you buy from us, the reference is stated in our Terms: the country and city of an IP are the ones ipinfo shows. A site that reads another database can show another city for the same address.

The score is not the type

On Scamalytics, Google's datacenter address scored 0 and the Comcast home provider address scored 16. Both are rated low risk, and the order is the opposite of what many people expect. The score measures suspicious traffic seen from an address and its network; it does not say whether the address is a home line or a server. That is a separate field: Scamalytics answered datacenter No for Comcast and Yes for Google and Hetzner, and ip-api's hosting flag split the four addresses the same way, false for the two home providers and true for the two datacenters.

So read a risk checker in two steps: the type fields first, the score second. Our own IP fraud score check is laid out that way, with the score, the flags and five blacklists in one view; in its published sample a Comcast resolver also scored 17 while two Tor exits scored 0.

IPQS and Scamalytics: why the numbers differ

Our run read the four checkers in the table above; what follows on IPQS is how our own product pages and Terms describe it. Both services give a fraud score where lower is better, and the similarity ends there. Scamalytics states on each result how it got the number: for the Comcast address it had no visibility of that single IP and applied the score of the operator's network. IPQS returns its score with connection and abuse signals, and its request settings, strictness among them, can affect the assessment. A fair comparison over time keeps the same checker and the same settings, and records the date.

That is also why a plan names its checker. Where one of our plans says Fraud score 0, each IP is checked on IPQualityScore right before delivery and has a score of 0 at that moment, as our Terms state; on our low fraud score ISP proxies, low fraud score is the overall description of the IPs across checkers, not a promise of zero on every one. Scores move with time and with use, so check the one your application depends on.

One rule: look the address up from outside

Scoring sites log the visits they receive. If you open one in a browser that is using the proxy, the proxy's address becomes one more visitor in their data. Read the exit address first, then look it up from another connection: put the address in the checker's URL, or use our tools, which look it up from our servers.

# 1. read the exit address through the proxy
curl -s -x http://USER:PASS@HOST:PORT https://api.ipify.org

# 2. look it up from your own connection
curl -s https://ipinfo.io/203.0.113.10/json

Replace the example address with your exit. The fields of the answer are explained one by one in how to check a proxy IP.

Which checker to use for a proxy

  • Is the IP in the right country? ipinfo, then the proxy location checker to see six more databases. We read one of our own ISP exits that way on 20 November 2026: 7 of 7 databases placed it in the United States.
  • Whose network is it? The ASN on ipinfo or ip-api. On city and carrier plans you pick the carrier by its ASN before you pay.
  • Home line, mobile or server? The type fields: mobile and hosting on ip-api, datacenter on Scamalytics.
  • How does a risk system rate it? The scorer your target relies on, with fixed settings. Our mobile proxy fraud score test shows what one network scored on 55 exits.

Then pick the network for the job: residential proxies for home IPs in many countries, mobile for carrier IPs, ISP proxies when one address has to stay the same for weeks.

Sources & further reading

FAQ

Quick answers on ip checkers compared.

Something else? Ask us

What is the difference between IPQS and Scamalytics?

Both return a fraud score, each from its own traffic, lists and model, so the same IP can score differently. Scamalytics may apply the score of the operator's network, as it did for one address in our run; IPQS results can change with its request settings, strictness among them, so compare with fixed settings.

Is ipinfo or IP2Location more accurate?

For the country they agree: in our run of 20 November 2026 four checkers gave the same country in 16 of 16 answers. The city is each database's estimate and can differ; for proxies bought from us, the reference in our Terms is ipinfo.

What does a Scamalytics fraud score mean?

An estimate, from 0 to 100, of how much traffic from an address or its network looks fraudulent. It is not the address type: in our run a datacenter address scored 0 and a home provider address scored 16.

Why do IP checkers show different cities for the same IP?

Each database builds its city data from its own sources. One address in our run came back as four cities on four checkers because it is announced from many places at once. The country matched on all four.

Should I open a fraud score site through my proxy?

No. Those sites log every visit, so the proxy's address becomes part of their data. Read the exit address, then look it up from another connection or with a tool that queries from its own servers.

Check an address on seven databases at once

The free proxy location checker and IP fraud score check look every address up from our servers.

Related reading