Fractal OSI Overview
Fractal OSI is the Ontological Semantic Information Modeler — a visual platform where domain experts, architects, and engineers define the ontological types, semantic relationships, and recursive structure of any domain without limits or predefined schemas.
Every complex system generates diagrams. The problem is that most diagramming tools are optimized for visual clarity, not semantic meaning. A box is a box. An arrow is an arrow. The tool has no idea — and neither will anyone reading it six months later — whether that box represents a department, a database table, a product, or a human being.
Fractal OSI is a different kind of tool. Every object placed on the canvas has a formal ontological type. Every relationship drawn between objects is a typed entity with attributes and identity. The platform enforces the rules of your information model as you build — not as a constraint, but as a guarantee of meaning.
The Three Words That Define It
The name OSI is intentional and precise:
| Letter | Meaning | What It Means in Practice |
|---|---|---|
| O | Ontological | Every object is assigned a formal ontological type — Entity, Role, or Act — before it can exist on the canvas. The platform models being, not just appearance. |
| S | Semantic | Every relationship drawn carries a defined semantic class — Participation, RoleLink, ActRelationship, or ScopeLink. Connections are not lines; they are typed objects with meaning. |
| I | Information | The platform operates at the conceptual information architecture level — above databases, above programming languages, above any specific technology stack. |
And Fractal — because any model you build can become a component (a FractalSet) inside a larger model, to infinite depth. Recursive composition is not a feature added on top; it is the architectural foundation.
What You Build in Fractal OSI
Fractal OSI is used to produce:
- Conceptual information models of any business domain — healthcare, finance, government, technology, logistics
- Organizational structures as navigable, semantically typed networks — not flat org charts
- Compliance and audit chains where every control, policy, role, and evidence item is formally typed and connected
- AI knowledge bases —
.fractalfiles that serve as typed relational graph inputs for LLM queries and code generation - Architecture specifications that compile progressively to TypeScript interfaces, API routes, and database schemas
Not a Diagramming Tool
Fractal OSI is frequently compared to draw.io, Lucidchart, or Miro. It is not in that category. The distinction is precise:
- A diagramming tool produces pictures that communicate to humans. The shapes have no formal type. The connections carry meaning only in the viewer's interpretation.
- Fractal OSI produces information models that communicate to both humans and machines. Types are enforced. Relationships are validated. Exported
.fractalfiles preserve full semantic structure.
Similarly, it is not UML — which is a software-paradigm modeling language — or an ER tool — which begins at the physical database layer. Fractal OSI operates at the conceptual information layer, above all implementation choices.
See It as a Model, Not a Picture
The Revenue Leakage information model below is a live canvas, not an image. Every stage is a typed Entity, every transition a typed Act, and the friction metadata lives on the links between them. Zoom in and click the (+) icon on a node to drill into its nested structure.
Where to Start
- New to Fractal OSI? → Quick Start: Your First Model
- Want the conceptual foundation? → The Three Principles
- Comparing tools? → Fractal OSI vs. Other Tools
- Looking for a working example? → Start from a Solution

