Domain Intelligence Schema

Terraform for Agentic AI Systems
v1.6.0
Schedule a Meeting
View Documentation GitHub Repository

What is DIS?

The Domain Intelligence Schema is a declarative framework for modeling and deploying agentic AI systems in enterprise environments. Think of it as the BNF for business units — a formal grammar for composing valid domain models. DIS captures domain structure, behavioral rules, execution bindings, and governance in a vendor-neutral format that compiles to any runtime platform including LangGraph, CrewAI, AWS Bedrock Agents, and Microsoft AutoGen.

Key Features

23 Core Constructs

A complete vocabulary for modeling entities, roles, behaviors, execution bindings, and governance in agentic AI systems.

7 Canonical Modes

Atomic verbs — CREATE, READ, UPDATE, DELETE, INITIATE, RESPOND, NOTIFY — that force proper entity discovery and prevent scope creep.

Zero Vendor Lock-in

Compile to LangGraph, CrewAI, AWS Bedrock Agents, Microsoft AutoGen, or any custom platform. No proprietary references.

Quick Start

Reference schemas directly in your IDE, validate with AJV, or generate types for your language.

IDE — $schema reference
{
  "$schema": "https://schemas.domainintelligenceschema.org/dis/1.6.0/DISDossier.schema.json",
  "dossierId": "customer-service-v1",
  "name": "Customer Service Domain",
  "disSpecificationRef": "1.6.0"
}
Validation — AJV
# Validate a dossier against the CDN-hosted schema
ajv validate \
  -s https://schemas.domainintelligenceschema.org/dis/1.6.0/DISDossier.schema.json \
  -d my-dossier.json
Type Generation — TypeScript
json-schema-to-typescript \
  https://schemas.domainintelligenceschema.org/dis/1.6.0/Entity.schema.json \
  > types/Entity.ts

Schema Categories

Foundation

2 schemas

CommonEnums (31 controlled vocabularies) and DISDossier (root container and manifest).

Identity & Structure

5 schemas

Entity, Role, Application, Endpoint, and Modes — the core nouns and interfaces of your domain.

Behavioral Grammar

2 schemas

AgenticTriplet (Entity + Mode + Role) and ModeOfInteractionRegistryEntry for behavioral composition.

Matrices

4 schemas

AccessGateMatrix, EntityModeMatrix, RelationshipMatrix, and TripletFunctionMatrixEntry for cross-cutting rules.

Supporting Constructs

4 schemas

FunctionCatalogEntry, KnowledgeDocument, TagDefinition, and ValidationRule for execution and metadata.

Governance

6 schemas

ValueEngineering, DossierComparison, PrivacyManifest, Telemetry, Marketplace, and ChangeManagement.