Skip to main content

Ontology Backend

The Ontology Backend manages the semantic model — the structured definition of business entities, metrics, dimensions, and their relationships. It serves this model to Cube and to AI features that need business context to generate accurate SQL or explanations.

Git-Backed Model

The ontology is stored in a Git repository. The backend syncs from the repo on startup and on demand, and can optionally push model changes back to the repo when edits are made through the API.

ConfigDescription
git_repo_urlSSH URL of the ontology Git repository
git_branchBranch to track
git_push_enabledAllow the backend to commit changes back to the repo
git_credentials_enabledMount SSH credentials for private repos
git_ssh_private_keyBase64-encoded SSH private key

Integrations

ConsumerWhat it gets
CubeThe data model — schema definitions, measures, dimensions, joins. Cube fetches from the ontology backend at startup and on version change.
Text-to-SQL / AI agentsBusiness-level context that improves SQL generation — entity names, metric definitions, and relationship metadata

Authentication

Keycloak OIDC. Can be toggled off for development.

Go Deeper

  • Cube — the semantic layer engine that consumes the ontology model
  • Text-to-SQL — AI feature that uses ontology context for query generation