MSP-1 Specification
site
The site term defines the website as a top-level semantic entity in MSP-1. A site represents the global identity, canonical domain, intent, authority, and protocol context under which all pages and resources operate. It enables AI agents to understand the scope, ownership, trustworthiness, and structural consistency of an implementation.
1. Purpose
The site term provides a foundational semantic anchor for MSP-1
implementations. AI agents rely on site-level metadata to:
- Identify the canonical domain and authoritative identity of the website.
- Infer global trust signals such as provenance, authorship, and compliance.
- Establish context inherited by all subordinate pages.
- Apply the correct protocol version and verification rules.
- Detect multi-site or multi-domain structures.
Site metadata acts as the “root of authority” for the entire MSP-1 ecosystem.
2. Normative definition
A site represents the canonical identity of a website within MSP-1.
A site MUST:
- Define a globally unique
idfor the site. - Provide its canonical domain (e.g.,
"msp-1.org"). - Declare site-level protocol metadata when applicable.
- Serve as the parent context for all pages.
- Remain stable unless a complete rebrand or domain shift occurs.
Sites may include extended metadata for trust, authority, or verification.
3. Required fields
- id — stable identifier for the site.
- name — human-readable site name.
- url — canonical homepage URL.
Recommended fields include:
- description — summary of the site mission and purpose.
- intent — the site’s high-level goals.
- provenance — origin and authorship lineage.
- protocol — the MSP-1 version and compatibility range.
- authority — organizational or individual ownership.
- trust — site-level trust indicators or verification requirements.
Site definitions SHOULD be published in .well-known/msp.json for discovery.
4. AI interpretation rules
- AI agents MUST treat the site object as the root semantic entity of the domain.
- Site-level metadata SHOULD cascade downward unless overridden by page-level declarations.
- Conflicting site metadata MUST be treated as a severe trust signal failure.
- AI SHOULD use site identity to merge identical resources across subdomains or mirrored URLs.
- If site metadata is missing, AI SHOULD infer basic identity from domain and structure, with reduced confidence.
Site metadata enables consistent, domain-wide interpretation across all MSP-1 resources.
5. Relationship to related MSP-1 terms
- page — each page belongs to a site.
- section — inherits context indirectly from the site.
- protocol — site-level protocol declarations define global rules.
- authority — identifies ownership of the site.
- provenance — documents origin and content lineage.
- trust — site-level trust indicators influence overall interpretation.
- version — site may declare overall version of its MSP-1 implementation.
6. Examples
Minimal site declaration:
{
"site": {
"id": "msp-1-org",
"name": "MSP-1 — Mark Semantic Protocol",
"url": "https://msp-1.org/"
}
}
Site declaration with recommended metadata:
{
"site": {
"id": "msp-1-org",
"name": "MSP-1 — Mark Semantic Protocol",
"url": "https://msp-1.org/",
"description": "An open, AI-first metadata protocol for machine-readable trust and clarity.",
"intent": "Provide a universal structured semantic layer that AI agents can reliably interpret.",
"provenance": { "type": "original" },
"authority": { "id": "mark-johnson", "role": "owner" },
"protocol": "MSP-1",
"version": "1.0.0"
}
}
Site declaration in a .well-known/msp.json context:
{
"protocol": "MSP-1",
"version": "1.0.0",
"site": {
"id": "msp-1-org",
"name": "MSP-1 — Mark Semantic Protocol",
"url": "https://msp-1.org"
},
"supportedVersionRange": "MSP-1.0.x"
}
7. Conformance
A resource conforms to the MSP-1 site specification when:
- It provides a stable, canonical identity for the site.
- The site metadata is truthful and consistent across all resources.
- All subordinate pages inherit or align with site-level declarations.
- All normative requirements for site semantics are satisfied.