AI Domain Data Standard — Specification Snapshot (v0.1.1)
The AI Domain Data Standard defines a simple, authoritative JSON document that any domain owner can publish to describe their identity to AI systems. The design is intentionally narrow—similar in spirit to RFC 9116 (security.txt)—with a single canonical record that requires no SaaS dependencies and no vendor lock-in. Version v0.1.1 is scoped strictly to identity fields and establishes the baseline structure and retrieval process for future extensions.
Scope of v0.1.1
Identity fields included in this draft:
- name
- description
- website
- logo (optional)
- contact
- entity_type (optional)
- jsonld (optional)
v0.1.1 (backward-compatible with v0.1) is published as an Informational Draft. Future versions will follow the same versioned-document and public change-tracking model used here.
Full Specification
Read the complete v0.1.1 specification document →
The full specification includes detailed field definitions, examples, resolution rules, and interoperability guidelines.
Resources
- Specification Repository: https://github.com/ai-domain-data/spec
- Download Schema (v0.1.1): schema-v0.1.json
- GitHub Action: ai-domain-data-validate-action
Resolution Order
- Primary Source: Retrieve
https://<domain>/.well-known/domain-profile.json. - Secondary Source: Query the DNS TXT record at
_ai.<domain>. - Authoritative Rule: When both sources exist, the HTTPS document served at
/.well-known/domain-profile.jsonis authoritative.
Base64 & Canonicalization Requirements
When publishing the DNS TXT variant, publishers MUST:
- Encode the UTF-8 JSON using the standard Base64 alphabet (no URL-safe variant).
- Not insert line breaks.
- Split TXT values at ≤255 characters.
For both HTTPS and DNS variants:
- Omit optional fields when unused.
- Do not include additional, undefined properties.
- Preserve URLs exactly as provided (including trailing slashes).
- Resolvers MUST treat field values as canonical and SHOULD NOT rewrite whitespace or the organization of the JSON object aside from parsing.
Document Structure
- Introduction & Terminology
- JSON Schema and Field Definitions
- Publishing Instructions (HTTPS + DNS TXT)
- Resolution & Validation Rules
- Canonicalization Requirements
- Security Considerations
- Appendix: Outreach Briefs & Ecosystem Roadmap
Contributing
All discussion, proposals, and revisions are tracked publicly in the specification repository: https://github.com/ai-domain-data/spec.
Submit issues or pull requests referencing the affected section or schema version. Versioned drafts and change logs are maintained transparently.