Data model

Your data model is the schema layer of your context lake - it defines what the things your organization cares about are called, what properties describe them, and how they relate to one another.

Combined with data ingestion, which populates that schema with real, live entities, the data model turns your software catalog into an organizational semantic layer: something AI agents, workflows, and dashboards can query and act on with full context.

Blueprints, properties, and relations

Blueprints are the basic building block of your data model - each one represents an asset in your organization, such as a microservice, a Kubernetes cluster, or a cloud account. A blueprint is made up of properties, the customizable fields that hold the data ingested for that asset, and connected to other blueprints through relations, the logical connections between them (for example, a service depends on a package, or a deployment runs in an environment).

Port's plug & play integrations come with predefined blueprints, properties, and relations out of the box, but you can freely edit them or create your own to match your organization's structure.

Editing the data model can be done from the UI, Port's API, or simply by describing the change in natural language through Port's AI assistant or an agent connected via the Port MCP server.

Define your ontology

A schema alone tells Port what fields exist. An ontology tells it what those fields mean - through descriptions on blueprints and properties, semantic relation titles, and well-chosen property types. This is what lets AI agents reason about your catalog correctly instead of just reading raw values.

Manage your data model as code

If you prefer to version-control your data model alongside your infrastructure, Port's data model can be defined and managed using Terraform or Pulumi, in addition to the UI, API, and Port AI.