Ontologies
Ontologies
An ontology is a typed model of a domain: the kinds of things in it, the properties each one has, and the named links between them. Ontologies let people and AI agents work from one shared model instead of many tables. In an industrial operation, an ontology names each asset, record, event, site and person the operation holds, anchors each one in the system that stores its data, and joins them by typed links. CodeNinja Atoms has published 44 ontologies of real operations across 12 sectors in the United States (19), Saudi Arabia (13), Pakistan (11), Germany (1): 643 objects and 560 links, each downloadable as hyper-ontology/1 JSON under CC BY 4.0.
What an ontology contains
Objects of a few kinds, each with properties, status values and links. The counts are read from the 44 published ontologies.
| Kind | What it is | Objects |
|---|---|---|
| record | what the operation writes down: work orders, tickets, findings, calibrations | 180 |
| asset | equipment and other physical things the operation owns and maintains | 155 |
| event | something that happens at a time: an alarm, a fault, a mission | 90 |
| actor | a role that acts on the operation: a desk, a crew, an officer | 47 |
| site | a place the operation runs: a station, a field, a centre | 46 |
| measure | a reading, sample or score taken from the operation | 38 |
| document | a report, letter or dataset the operation keeps | 35 |
| person | a named job a person holds: the one who reads, decides or approves | 20 |
| system | a software or control system the model names as an object in its own right | 16 |
| material | stock that moves through the operation: parts, wafers, containers | 15 |
Ontology examples by sector
Every published ontology belongs to one real operation. Open a design to read its object model, or download it as JSON.
- Industrial ontology examples: every published object model, browsable and downloadable
- Industrial data ontology: what an industrial data ontology holds, kind by kind
Ontologies for AI agents
An agent that reads an ontology reasons over named objects instead of raw rows, and every action it proposes names the object it touches. The published ontologies load with the open loader, are served to agents by the MCP server, follow the hyper-ontology/1 format, and stand up as a living ontology in Hyper Ontology.
Built from the published designs only: each paper's At a glance table, its object model and the dataset. Designed on Praxis; object models load into Hyper Ontology. Also: Physical AI · Industrial AI · Self-improving systems · sovereign AI. Updated 2026-11-14.
Want one of these designs for your own operation? Talk to us.