IPinfo vs MaxMind: how to actually pick one

Quick note on where we stand: we sell fraud scoring, not IP data. Neither of these is our competitor, and we have nothing to gain from you picking either one.

So this is not a scored table with a winner.

The question that settles it#

Most people arrive asking which is more accurate. That is the second question.

The first one is: can you afford a network call on every request?

Scoring a signup? Then yes. One extra API call on a form submission is nothing.

Enriching every request at your edge, every row in a log pipeline, every analytics event? Then no. Millions of lookups a day over the network costs you latency and money you do not need to spend, because this kind of data ships as a file you can search locally in microseconds.

MaxMind is built around that file model. IPinfo offers downloads too, but the API is the front door.

Answer that one question and your shortlist usually drops to one.

What each is best at#

MaxMind has been mapping IP addresses since before most of these companies existed. Two things follow from that: the free tier is genuinely useful rather than a crippled demo, and the downloadable database story is mature in a way that is hard to catch up with.

They also sell minFraud, which is a different thing: a risk score, from about $0.005 a lookup, with no minimum. If you want risk rather than facts, look at that. Pay as you go with no minimum is unusually friendly at low volume.

IPinfo is better at context. Which network, which company, which mobile carrier, plus VPN and proxy detection on the higher plans. If you want a rich description of an address rather than a number about it, that is the shape you want.

What neither of them can tell you#

Both answer questions about an address. Neither answers a question about a person.

An IP cannot tell you that the same person opened forty accounts last week under forty different addresses. Or that the email domain was registered eleven minutes ago. Or that this device has signed up four times today.

None of that is in an IP address, and no vendor can put it there.

There is a bigger limit too, and it catches people out: the address is not the person.

Mobile networks put thousands of unrelated customers behind a single IP. So do offices and universities. Block on address reputation and you punish geography rather than behaviour, and the customers it costs you are mostly mobile users in exactly the countries where that sharing is universal.

About proxy detection#

This is what most people are really shopping for, so here is the straight version.

Neither vendor publishes proxy-detection numbers you can independently check. Treat every claim on both sides as marketing until you have tested it on your own traffic, which costs you a week of logging both and comparing against what you already know happened.

Two things are true of both:

Data centre addresses are easy. They are announced and allocated to hosting companies. Everyone catches them.

Residential proxies are hard. They route through real people's home broadband. An address belonging to a genuine subscriber does not look like a data centre, because it is not one. This is where every IP vendor quietly falls apart, ours included.

So if your abuse still arrives from data centres, either vendor catches it and you should buy the cheap one. If it does not, no IP vendor is going to be your answer, and it is worth knowing that before you spend a quarter integrating one.

Try the free tier first#

MaxMind's free tier needs an account and gives you the same lineage as the paid databases. It is less precise on cities and updated less often.

For the two things fraud work usually needs, a country and a network, it is normally enough.

Worth trying before you pay for anything, because it tells you what the paid version has to beat. A vendor demo cannot tell you that. Your own traffic can.

A sensible setup at low volume#

Use a downloaded file for the lookup you do on every request. It costs nothing per call and adds no delay. Then spend money only where money is involved.

// every request: local file, microseconds, no cost per call
const { country, asn } = geoip.lookup(req.ip);

// only events that move money get the expensive path
if (event === "cashout" || event === "payment_attempt") {
  const r = await kaidn.score({ event, user_id, ip: req.ip, email, device_id });
  if (r.verdict === "block") return hold(r.reason_text);
}

What you should not do is buy two products that answer the same question and average them. Two vendors disagreeing gives you a tiebreak problem, not better data.

Where we sit#

We are not an alternative to either for IP data, and we would rather say that than imply otherwise.

The difference is what comes back. Both of these return facts about an address. We return allow, review or block about a person, with the IP as one input beside the email, the device, and how often the same thing is repeating.

If you only ever read the IP part of our answer, you have bought an expensive IP lookup. Buy theirs instead.

Want to see what an address returns without signing up for anything? The free IP checker runs it with no account.