Visual editor (OntoViz)
Drag-and-drop entities, relationships and inheritance hierarchies on an interactive canvas.
OntoBricks features
Everything OntoBricks does, at a glance — from designing an ontology to serving a governed, AI-ready knowledge graph. The whole pipeline runs on the Databricks Platform.
OntoBricks takes your Unity Catalog tables through five stages — each one automated, each one editable.
Design
Ontology in OWL
Map
R2RML to tables
Build
Materialize triples
Reason
OWL · SWRL · SHACL
Serve
GraphQL & MCP
Ontology design
AI Assistant
Reshape your ontology with natural-language commands — add entities, remove orphans, list relationships.
SWRL rules
Author inference rules in a graphical IF/THEN editor with live preview and raw-edit mode.
SHACL data quality
Six shape categories with Turtle round-trip and SQL compiled for triple-store execution.
OWL axioms & constraints
Cardinality, value restrictions, functional/transitive/symmetric, equivalence and disjointness.
Pitfalls detector
Catch 19 structural, logical and semantic modelling pitfalls right from the sidebar.
Data mapping
Visual mapping designer
Bind ontology classes and relationships to Databricks tables in an interactive designer.
AI-powered wizard
An LLM writes the SQL from your table metadata, column by column.
Attribute-level mapping
Multi-pass column matching with per-attribute include/exclude and partial-mapping detection.
Auto-Map
Batch-map every unmapped entity and relationship asynchronously, with progress tracking.
R2RML generation
W3C-compliant R2RML mappings generated automatically from the visual configuration.
SQL query testing
Test and preview your mapping queries with live results before you save.
Knowledge graph — build & explore
Every build materializes two layers: a governed Delta view in Unity Catalog, and a graph engine you pick per domain.
R2RML mappings
→
Knowledge Graph build
→
Delta view Unity Catalog · governance
Graph DB engine Lakebase · Lakehouse · Neo4j
Graph Explorer
A sigma.js WebGL viewer with search, depth control, filters and entity detail panels.
Cluster detection
Louvain, Label Propagation or Greedy Modularity communities, with color and super-nodes.
Graph analytics
Five centrality metrics computed over the full graph by a serverless Lakeflow job.
AI interpretation
An agent turns analytics into key findings, notable entities and recommendations.
Cohort discovery
Rule-based, explainable cohorts materialized as graph triples or Delta tables.
Class actions & virtual attributes
Run Unity Catalog functions on a node, or compute attribute values on demand.
GraphQL API
Auto-generated schema
Types, fields and typed relationships derived from your ontology at runtime.
Nested traversal
Query relationships in a single request instead of flat triple lists.
GraphiQL playground
An in-browser IDE per domain with autocomplete, docs explorer and history.
SDL introspection
A machine-readable schema endpoint for external tooling and codegen.
MCP & AI integration
Model Context Protocol
Serve the graph to LLM agents over Streamable HTTP and stdio transports.
Two-step workflow
list_domains then select_domain, then explore entities and types.
GraphQL via MCP
Agents introspect and query the typed GraphQL API through dedicated tools.
Per-domain policy
Choose which tools each domain publishes and how its context is surfaced.
Playground & clients
Auto-discovered in the Databricks Playground; works with Cursor and Claude Desktop.
Project & registry
Unity Catalog storage
Domains stored in UC volumes with the registry in Lakebase Postgres.
Version lifecycle
DRAFT → IN-REVIEW → PUBLISHED with reviewer sign-off and quorum.
Single-editor locking
Safe concurrent editing via a renew-only lease with auto-release.
Audit trail
One newest-first feed of ontology changes, review decisions and build runs.
Import / export
Bring in FIBO, CDISC, IOF, HL7 FHIR; export portable .obx domains.
Configurable branding
Set the app title, primary color, logo and default entity icon.
Platform & operations
Databricks-native
Browse Unity Catalog, run on SQL Warehouses, embed AI/BI dashboards.
Security
CSRF protection, secure session cookies, and role-based access control.
Performance
SQL connection pooling and a dedicated thread pool for blocking I/O.
Observability
Structured JSON logging and per-request timing for every call.
Deployment
One-command Databricks Asset Bundle deploy for the app and its MCP server.