Universal Ontology

You can deploy and run a configurable MCP server that helps you explore, validate, and build knowledge graphs from ontologies. This server provides intelligent semantic search, proactive schema guidance, component-based modeling, ASCII graph visualization, built-in SHACL validation, and controllable session management to maintain clean, accurate graphs.

How to use

To use the MCP server, start it locally and connect with an MCP-compatible client. You will load your ontology files, configure the embedding model, and then interact through the client’s commands to create entities, set properties, attach components, and validate or export your graph.

How to install

Prerequisites: ensure Python is installed on your system.

Clone the project to your workspace.

Install dependencies.

git clone https://example.com/seonwookim92/universal-ontology-mcp.git
cd universal-ontology-mcp

Set the ontology directory path you will use for your .ttl files.

export ONTOLOGY_DIR="/path/to/your/ontology/folder"

Additional sections

Configuration and runtime details are provided below. You can run the MCP server locally using Python and point it at your ontology files. The server’s environment can specify the directory of ontologies and the embedding model to use for semantic reasoning.

Environment variables shown in the examples include:

• ONTOLOGY_DIR: Path to your ontology folder containing .ttl files.

• EMBEDDING_MODEL: Name of the embedding model used for semantic search (for example, all-MiniLM-L6-v2).

You can connect using an MCP client with one of the provided configurations. The server supports local (stdio) execution where the client spawns the Python process and talks to it directly.

Here are two example configurations shown in code form. Use the one that matches your setup.

{
  "mcpServers": {
    "universal-ontology-mcp": {
      "command": "python",
      "args": ["/absolute/path/to/universal-ontology-mcp/main.py"],
      "env": {
        "ONTOLOGY_DIR": "/absolute/path/to/your/ontology/folder",
        "EMBEDDING_MODEL": "all-MiniLM-L6-v2"
      }
    }
  }
}

{
  "mcp": {
    "universal-ontology-mcp": {
      "type": "local",
      "command": ["python","/Users/.../main.py"],
      "environment": {
        "ONTOLOGY_DIR": "/Users/.../ontology",
        "EMBEDDING_MODEL": "all-MiniLM-L6-v2"
      }
    }
  }
}

Available tools

get_ontology_summary

Provide a quick high-level overview of the loaded schema.

search_classes

Perform weighted semantic discovery over classes in the ontology.

search_properties

Search properties with semantic weighting to find relevant predicates.

get_class_details

Return detailed usage instructions and connectivity rules for a class.

list_available_facets

Rank and list components suitable for complex data grouping.

create_entity

Create a new entity with automatic URI sanitization.

set_property

Set a property value on an entity.

attach_component

Attach a component to an entity or class in the graph.

remove_entity

Remove an entity from the graph in a surgical way.

reset_graph

Reset the entire graph for a fresh session.

validate_entity

Check an entity against SHACL constraints for validity.

visualize_graph

Preview the graph as an ASCII tree, with verbose mode available.

export_graph

Export and save the validated graph to a TTL file.