Namespaced Identifier (NSID)

A Namespaced Identifier (NSID) in the AT Protocol is a structured identifier used to reference lexicon schemas for records, XPRC endpoints, and other protocol components. NSIDs provide a hierarchal naming system that ensures uniqueness across the ATmosphere.

Structure

NSIDs follow a reverse domain-name pattern followed by a name segment:

tld.domain.name

For example, com.example.fooBar would be a valid NSID, with com.example being the domain name and fooBar being the name segment.

This structure ensures that organizations can create identifiers within their own controlled namespace, preventing naming conflicts across the protocol.

Syntax

NSIDs must adhere to specific syntax rules.

In the overall structure, the NSID must:

• Contain only ASCII characters
• Have at least 3 segments separated by periods
• Contain a maximum length of 317 characters

In the domain authority portion, the NSID must:

• Be in reverse domain-name order (e.g. com.example instead of example.com)
• Have at least 2 segments
• Contain a maximum of 253 characters (including periods)
• Contain only lowercase letters, digits, or hyphens (cannot start or end with hyphens)
• Must not start with a digit

In the name segment, the NSID must:

• Be 1-63 characters long
• Contain only letters and digits (case-sensitive)
• Must not start with a digit

Variations

NSIDs can be extended with:

• Fragments, using a hash symbol to reference specific definitions within a schema (e.g. com.example.defs#userView),
• Glob patterns, using an asterisk to match multiple NSIDs (e.g. com.atproto.*)

Authority and Control

The domain authority portion of an NSID implies ownership and control. Only those who control a domain should create NSIDs within that namespace. This decentralized approach to naming allows different organizations to independently develop and publish schemas without central coordination.