Skip to content
Unified Modelling Schema
Unified Modelling Schema

Unified Modelling Schema

A universal schema for data structure and meaning.
Get StartedView Specification

Model once. Use everywhere.

Section titled “Model once. Use everywhere.”

Unified Modelling Schema captures both data structure and meaning in one portable model. Use the same schema across relational, graph, document, and multi-model systems.

Start with the model

Define types, portable properties, keys, constraints, relationships, and their natural-language readings in one YAML document.

Keep the semantics

Fact type readings make the intended meaning explicit for people, generated documentation, and AI-assisted work.

Choose your target

Use the same UMS model when designing relational tables, property graphs, document structures, or a mixed estate.

Find your paradigm

The reference explains how UMS concepts translate for relational, graph, document, and multi-model practitioners.

Simple to read. Precise enough to implement.

UMS uses YAML to describe data structures together with their meaning. A model remains readable by people while providing a structured definition that software and AI can process.

Types:
  - Type: Person
    Labels:
      - Person
    PrimaryKey:
      - Person_Id
    Properties:
      - Name: Id
        DataType: Integer
        Constraints:
         - NOT NULL
         - UNIQUE
        FactTypeReadings:
         - Language: English
           Readings:
            - {Person} has {Id}
            - {Id} is for {Person}
      - Name: LoginName
        DataType: TextVariableLength
        Constraints:
         - NOT NULL
         - UNIQUE
        FactTypeReadings:
         - Language: English
           Readings:
            - {Person} has {LoginName}
            - {LoginName} is of {Person}


      - Name: Name
        DataType: Text

Relationships are First-Class Citizens.

Relationships are Types in their own right. They can be graph relationships or join tables, at the same time.

- Type: PersonLikesFilm
  Labels:
  Label: LIKES
  Source: Person
  Target: Film
  RelationshipAnnotation: {Person} likes {Film}
  PrimaryKey:
  - Film_Id
  - Person_Id
  Properties:
  - Name: Rating
    FactBasedName:
    Alias:
    DataType: TextVariableLength
    Length:
    Precision:
    Constraints:
    FactTypeReadings:
    Facts:
    IsArray:
    ArrayConstraints:
  - Name: Film_Id
    FactBasedName:
    Alias:
    DataType: Integer
    Length:
    Precision:
    Constraints:
    - NotNull
    FactTypeReadings:
    Facts:
    IsArray:
    ArrayConstraints:
  - Name: Person_Id
    FactBasedName:
    Alias:
    DataType: Integer
    Length:
    Precision:
    Constraints:
    - NotNull
    FactTypeReadings:
    Facts:
    IsArray:
    ArrayConstraints:
  Relationships:
  - Name: FilmPersonLikesFilm
    Label:
    Source: PersonLikesFilm
    Target: Film
    From:
    - Film_Id
    To:
    - Film_Id
    Embed:
    Cardinality:
    Readings:
    - {PersonLikesFilm} involves {Film}
    - {Film} is involved in {PersonLikesFilm}
    Facts:
  - Name: PersonPersonLikesFilm
    Label:
    Source: PersonLikesFilm
    Target: Person
    From:
    - Person_Id
    To:
    - Person_Id
    Embed:
    Cardinality:
    Readings:
    - {PersonLikesFilm} involves {Person}
    - {Person} is involved in {PersonLikesFilm}
    Facts:
  FactTypeReadings:
  - Language: Not Defined
    Readings:
    - {Person} likes {Film}
  Facts:
  UniquenessConstraints:
  IsRelationshipType: true

About Unified Modelling Schema

UMS provides a portable way to describe data structures, constraints, relationships, and natural-language meaning in one schema.

Portable

Use one schema across relational, graph, document, and multi-model platforms.

Semantic

Capture natural-language readings alongside structural definitions so meaning travels with the model.

Implementation-aware

Describe keys, properties, constraints, relationships, and target-specific details in a precise form.

AI-ready

Provide structured schema information and explicit natural-language predicates that AI systems can interpret.

One model. Many data paradigms.

Data platforms express many of the same modelling concepts in different ways. UMS provides a common schema in which those concepts can be described once and carried across technologies.

Relational

Describe tables, columns, keys, constraints, and relationships while retaining the conceptual meaning behind them.

Graph

Describe nodes, properties, labels, relationships, and graph-oriented structures using the same underlying model.

Multi-model

Bring relational, graph, document, and other representations together within a single schema.

Meaning travels with the model.

UMS can include natural-language predicates alongside structural definitions, giving people and AI systems an explicit description of what the model means.

This makes the schema useful for documentation, interpretation, automation, and AI-assisted modelling.

Person has Name

Name is for Person

Employee works for Company

Company employs Employee

What Unified Modelling Schema captures

UMS provides a common vocabulary for describing the structural, relational, semantic, and implementation-oriented parts of a data model.

Types and properties

Define entity types, value types, properties, datatypes, nullability, defaults, and other structural characteristics.

Keys and constraints

Capture identifiers, uniqueness, mandatory values, cardinality rules, and other constraints that govern valid data.

Relationships

Describe associations between types, including roles, direction, participation, and relationship-specific properties.

Natural-language readings

Attach explicit predicates and inverse readings so the meaning of the model remains available to people, documentation, and AI.

Platform mappings

Express how a conceptual model maps into relational, graph, document, and multi-model implementations.

Extensible metadata

Carry additional annotations and target-specific information without losing the shared model underneath.

One concept. Multiple implementations.

UMS lets a single model carry across different database paradigms while preserving the meaning of the original concept.

UMS

Person works for Company

The model defines the participating types, relationship, roles, constraints, and natural-language readings.

Relational

Person

CompanyId → Company.Id

The relationship can be implemented using relational keys and constraints.

Graph

(Person)-[:WORKS_FOR]→(Company)

The same relationship can be expressed directly as a graph relationship between nodes.

Meaning preserved

Person works for Company

Company employs Person

Schema + Semantics + AI

Your schema can explain itself.

UMS combines machine-readable structure with human-readable meaning, giving AI the context to understand what your data model represents.

Ask your model…

💬   What does Customer mean?

💬   How is Order related to Customer?

💬   Generate the relational schema.

💬   Show this model as a property graph.

💬   Explain this schema to a business analyst.

💬   Compare these two models.

From model to implementation.

A UMS schema can become the common source for the artefacts used throughout the data modelling and implementation lifecycle.

Source

UMS Schema

Types · Properties · Relationships
Constraints · Readings · Semantics

→

Generate and transform

→ Relational DDL

→ Graph schemas

→ Documentation

→ Data model visualisations

→ AI context

→ Platform-specific representations

Open · Portable · Semantic

Start modelling with UMS.

Explore the specification, create your first Unified Modelling Schema, and help shape a common language for data modelling.