Ontology
The ecosystem.Ai Ontology is how you discover and share the concepts the platform uses: projects, predictors, offer matrices, invocations, contact logs, and how they connect.
It is the semantic layer over:
- Workbench2 — configuration entities in
ecosystem_meta - ecosystem-server — modeling projects, diagrams, generation
- ecosystem-runtime — real-time
/invocationsand/response
Like an operational ontology elsewhere, it connects digital assets (Mongo collections, deployments, models) to their counterparts (customers, offers, channels) so products reason consistently about decisioning.
How to use this documentation
| Goal | Start here |
|---|---|
| Learn why ontology exists | Why create an Ontology? |
| Browse every class & property | Concept Index |
| Workbench meta concepts | Entities · full catalog |
| Runtime API & logging | Runtime · full catalog |
| Modeling server concepts | Server · full catalog |
| Fixed offer catalog shape | Offer Matrix |
| Open customer / feature docs | Customer Feature Lookup |
| Term lookup | Glossary |
| Industry domain seeds (public catalog) | ecosystem-domain-catalog on GitHub · live ontology.ecosystem.ai |
Modules at a glance
| Module | Namespace | Concepts |
|---|---|---|
| Entities | …/ontology/entities# | Project, Frame, Predictor, DynamicEngagement, OfferMatrix, Campaign, Contacts, DeploymentStep, … |
| Workbench | …/ontology/workbench# | Solution cases, AsyncJob, SyntheticDataRequest, RuntimeConsole |
| Server | …/ontology/server# | ModelingProject, Diagram, OntologyContext, generation jobs |
| Runtime | …/ontology/runtime# | InvocationRequest/Response, ContactLog, ResponseLog, plugins, Option |
| Platform | …/ontology/platform# | Umbrella + alignments |
Full IRIs use the base https://ecosystem.ai/ontology/. See Namespaces.
Two schema rules (critical)
- Offer Matrix rows are closed —
offer_id,offer_name,cost,price,category,static_offers - Customer / feature lookup is open — any document shape keyed by customer (or configured lookup key)
Published form
These pages are the published vocabulary. The site does not serve Turtle or SHACL files. Customers still describe their own open lookup documents, and Workbench can export an indicative ontology from a collection they own.
Last updated on