RELICA

Semantic Modeling for Next-Generation Applications

Ad Specmentis Foundation, LLC develops open-source software for precise, machine-readable semantic modeling. Built on Gellish Ontological Modeling Language, RELICA enables developers to create knowledge graphs where relationship semantics are defined within the graph itself, using the graph's own vocabulary.

The RELICA semantic modeling platform allows for the distributed processing of semantic relationships across knowledge domains using self-referential modeling patterns. It is designed to scale from simple ontologies to complex enterprise knowledge graphs, maintaining both local semantic consistency and global interoperability. Rather than hardcoding semantics in external schemas, RELICA embeds them directly in the graph structure, where they can be extended and refined using the graph's own vocabulary. This approach provides semantic stability where needed while allowing controlled evolution through meta-modeling patterns.

Project Updates

RELICA core architecture has stabilized after 2+ years of development and 3 major iterations. Platform now features production-ready microservice design with well-defined module boundaries.

Built on modernTypeScript stack withNestJS backend andReact frontend. Triple storage powered by Neo4j with Redis semantic caching layer.

Focus shifting from foundation to demonstrable value delivery. Open-source release planned following feature completion and selective collaboration phase.

Core Functionality

Self-Referential Semantics

RELICA goes beyond the simple expression of relationship to defining the explicit terms by which entities can relate. While traditional knowledge graphs leave relationship constraints external or undefined, RELICA embeds these rules within the graph itself - creating a self-contained system that informs what connections are valid and what they mean.

Foundation Ontology & Domain Extensions

  • • Bootstrap taxonomy providing core semantic primitives and relationship types
  • • Pre-built domain extensions for common modeling patterns:
    • ◦ Physical objects, states, and transformations
    • ◦ Events, occurrences, and causal relationships
    • ◦ Spatial and temporal relationships
    • ◦ Part-whole hierarchies and compositional structures
  • • Extensible foundation - use what fits, extend what doesn't
  • • Domain packages available for manufacturing, logistics, and enterprise systems

Semantic Data Management

  • • High-performance triple storage with graph-native architecture
  • • Semantic-aware caching for optimized query performance
  • • Comprehensive low-level API for direct fact access
  • • Advanced query facilities with semantic relationship awareness

Object-Semantic Mapping (OSM)

  • • Translation between raw triple representation and application objects
  • • Domain modeling framework for representing objects, states, events, and their relationships across space and time
  • • Pre-built conceptual patterns for common modeling scenarios (part-whole, cause-effect, spatial, temporal)

Semantic Environment Management

  • • Load/unload triples and objects into semi-persistent user-linked subgraph collections
  • • Dynamic working spaces for knowledge modeling

Data Integration

  • • Bulk load/unload of fact triples via XLS/CSV transport
  • • Format-preserving data exchange

AI-Powered Capabilities

  • • Integrated AI modeling assistance in administrative interface
  • • Programmatic AI services for semantic tasks (NER, relationship extraction, entity resolution)
  • • Knowledge graph exploration and pattern discovery

Platform Infrastructure

  • • User authentication
  • • Token-protected comprehensive API surface
  • • Administrative user interface

Use Cases

Living Domain Models

  • • Build actual models of your world, not just lists of things
  • • Define how things relate, change, and affect each other
  • • Applications that understand your domain like you do
  • • Stop flattening reality into rows and columns

Semantic Computing

  • • Programs that work with concepts, not just data fields
  • • AI that understands what things are, not just their labels
  • • Systems that know why things connect, not just that they do

Knowledge Environments

  • • Work in spaces where meaning is built-in
  • • Every connection carries its own explanation
  • • Explore implications, not just query records

Community

Documentation

Comprehensive documentation and examples available atcorpus-relica.github.io

Source Code

Open source development on GitHub atgithub.com/corpus-relica