Ontology MCP

Ontology MCP is a Model Context Protocol (MCP) server that connects GraphDB's SPARQL endpoints and Ollama models to Claude. This tool allows Claude to query and manipulate ontology data and leverage various AI models.

Key Features

SPARQL related functions

• Execute SPARQL query ( mcp_sparql_execute_query )

• Execute SPARQL update query ( mcp_sparql_update )

• List repositories ( mcp_sparql_list_repositories )

• Query the graph list ( mcp_sparql_list_graphs )

• Get resource information ( mcp_sparql_get_resource_info )

Features related to the Ollama model

• Run the model ( mcp_ollama_run )

• Check model information ( mcp_ollama_show )

• Download model ( mcp_ollama_pull )

• Get model list ( mcp_ollama_list )

• Delete model ( mcp_ollama_rm )

• Chat completion ( mcp_ollama_chat_completion )

• Check container status ( mcp_ollama_status )

OpenAI related features

• Chat completed ( mcp_openai_chat )

• Create image ( mcp_openai_image )

• Text-to-speech ( mcp_openai_tts )

• Speech-to-text ( mcp_openai_transcribe )

• Generate embedding ( mcp_openai_embedding )

Google Gemini related features

• Generate text ( mcp_gemini_generate_text )

• Chat completion ( mcp_gemini_chat_completion )

• Get model list ( mcp_gemini_list_models )

• ~~Generate images ( mcp_gemini_generate_images ) - Using Imagen model~~ (currently disabled)

• ~~Generate videos ( mcp_gemini_generate_videos ) - Using Veo models~~ (currently disabled)

• ~~Generate multimodal content ( mcp_gemini_generate_multimodal_content )~~ (currently disabled)

Note : Gemini's image creation, video creation, and multimodal content creation features are currently disabled due to API compatibility issues.

Supported Gemini Models

HTTP request functions

• Execute HTTP requests ( mcp_http_request ) - communicate with external APIs using various HTTP methods such as GET, POST, PUT, DELETE, etc.

Related MCP server: RDF Explorer

Get started

1. Clone the repository

git clone https://github.com/bigdata-coss/agent_mcp.git
cd agent_mcp

2. Run the GraphDB Docker container

Start the GraphDB server by running the following command from the project root directory:

docker-compose up -d

The GraphDB web interface runs at http://localhost:7200 .

3. Build and run the MCP server

# 의존성 설치
npm install

# 프로젝트 빌드
npm run build

# 서버 실행 (테스트용, Claude Desktop에서는 필요 없음)
node build/index.js

4. Import RDF data

Go to the GraphDB web interface ( http://localhost:7200 ) and do the following:

1. Create a repository:

   • “Setup” → “Repositories” → “Create new repository”

   • Repository ID: schemaorg-current-https (or whatever name you want)

   • Repository title: "Schema.org"

   • Click "Create"

2. Get sample data:

   • Select the repository you created

   • “Import” → “RDF” → “Upload RDF files”

   • Upload an example file to the imports directory (e.g. imports/example.ttl )

   • Click "Import"

Note : The project includes example RDF files in the imports directory.

5. Setting up Claude Desktop

To use Ontology MCP in Claude Desktop, you need to update the MCP settings file:

1. Open the Claude Desktop settings file:

   • Windows: %AppData%\Claude\claude_desktop_config.json

   • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

   • Linux: ~/.config/Claude/claude_desktop_config.json

2. Add the following settings:

{
  "mcpServers": {
    "a2a-ontology-mcp": {
      "command": "node",
      "args": ["E:\\codes\\a2a_mcp\\build"],
      "env": {
        "SPARQL_ENDPOINT": "http://localhost:7200",
        "OPENAI_API_KEY": "your-api-key",
        "GEMINI_API_KEY" : "your-api-key"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

IMPORTANT : Change the path in `args' to the actual absolute path to your project build directory.

3. Restart Claude Desktop

License

This project is provided under the MIT License. See the LICENSE file for details.