mcp-knowledge-graph

The mcp-knowledge-graph server is designed to enable persistent memory for the Claude AI model by utilizing a local knowledge graph. This fork of the Model Context Protocol (MCP) focuses on local development, allowing you to efficiently store and retrieve contextual information, enhancing the capabilities of your applications powered by Claude. By using this server, you can effectively manage and query knowledge with ease, tailored to your specific development needs.

Once you are connected to the mcp-knowledge-graph server, you can interact with it using various commands to store and retrieve data. You can issue queries to add new entries to your knowledge graph or fetch existing information using structured queries that leverage the graph’s capabilities. Best results are achieved with queries focused on specific contexts or topics relevant to your application, allowing Claude to draw from the most pertinent data during its operations.

To install the mcp-knowledge-graph server, ensure you have the following prerequisites:

• Node.js (version 14 or higher)

Option A: Quick start with npx
Since the NPM package is not specified, you can clone the repository directly:

git clone https://github.com/shaneholloman/mcp-knowledge-graph.git
cd mcp-knowledge-graph
npm install

Option B: Global install alternative
If you prefer a global installation approach after cloning, you can link the package:

npm link

This way, you can access the server globally from any directory.

When configuring the mcp-knowledge-graph server, ensure that you set any necessary environment variables for your local setup, such as paths to your data storage. A common gotcha is not having the appropriate permissions set for your local environment, so check that your user has the necessary access rights for file I/O operations.