What we publish

The purpose of the IRCStats project is to publish general statistics on IRC usage as a whole. We publish statistics on IRC server software, server features (CAP and TLS usage), and services packages. We publish overall numbers and percentages.

We do not publish statistics tied to a specific network. In particular we do not publish server names, network names, nick names or channel names. There are other websites for that.

This project is about researching overall IRC usage. The only exception is that the collected information (Dec'22 onwards) may also be used by us to contact you about possible security issues of your server. For example, if it seems to be running a vulnerable version of certain specific IRC Server software. We will not share any identifiable collected information with vendors (no server names, no e-mail addresses, etc), but can send a notification on their behalf or on our own initiative. Important: do not depend on such notifications in any way, you are responsible for the security of your own server!

What we do

Every year we scan the entire Internet (the IPv4 address range) for IRC Servers on port 6667 and 6697. Since Dec'22 we also scan IPv6 addresses on these ports. Because scanning the full IPv6 address space is impossible, we use the IPv6 hitlist and our own collected DNS information for a list of IPv6 IP's to scan.

Our current scanner bot sends the following commands: NICK, USER, CAP, VERSION, ADMIN, MODE, MOTD, PING, PONG, MAP, LINKS, LUSERS, MODULE, INFO, JOIN 0, QUIT. The output of these commands is saved and later used to produce general statistics. Only the general statistics are published.

As you can see we try to 'keep the noise down' by only connecting once per year. We also gather only minimal data. In particular we do not gather the list of channels (LIST) nor the list of users (WHO).

Why we do it

There's currently no reliable information on actual IRC usage on the Internet.

Can I opt-out?

Yes, you can opt-out here.