NodeInfo is a small metadata protocol used by many decentralised social servers to publish information about the software they run and the community they host. It can expose details such as the server software and version, supported protocols, usage counts and whether registrations are open.
How NodeInfo discovery works
Discovery starts at /.well-known/nodeinfo. That document points to one or more NodeInfo schema URLs, normally identified by version. A client follows the advertised link and receives JSON describing the server. NodeInfo 2.1 includes fields for software name and version, supported protocols, services, open-registration status, usage statistics and optional metadata. Because the discovery document and schema document are separate, a server can advertise the versions it supports without hard-coding one URL convention into every client.
Why instance metadata is useful
NodeInfo makes it possible to build instance directories, network statistics, compatibility tools and research datasets without scraping each project's HTML. A crawler can identify that one host runs Mastodon, another runs Misskey and another runs a different ActivityPub implementation even though their public websites look completely different. It is especially useful in the Fediverse because federation is a property of thousands of independently administered servers rather than one central service with a single API.
Why the numbers need context
The numbers in NodeInfo should not be treated as audited measurements. Implementations can calculate users, posts and activity differently, and administrators can omit optional fields. NodeInfo also describes a server; it does not enumerate every remote server with which that server federates. Support is widespread in parts of the Fediverse but not universal across every ActivityPub implementation or decentralised protocol.