Skip to main content

How to Install the InfraNodus MCP Server

Dmitry Paranyushkin
  • Updated

You can use the InfraNodus MCP server in order to add knowledge graph, GraphRAG, and text analysis capabilities to your favorite LLM workflows and AI automation tools. MCP is a protocol that acts as a layer between your software (e.g. ChatGPT) and the InfraNodus API that provides the model with various tools that it can combine for agentic workflows. The InfraNodus MCP server is open source, which means that you can add your own tools and calls if you deploy it locally.

In most cases, using the InfraNodus MCP server is an easier option than using the InfraNodus API, so you might want to start with it if you'd like to get programmatic access to InfraNodus' capabilities and data.

There are multiple ways to deploy the InfraNodus MCP server. The best way depends on where you intend to run the server and whether you want to customize it to your particular needs.

 

Deploying to a Cloud Service (Claude Web, ChatGPT, n8n, Dify)

The easiest way to deploy the InfraNodus MCP server to a cloud web service is to use our HTTP Streamable endpoint:

https://mcp.infranodus.com

Copy this URL and  add it to your Claude Web or ChatGPT via Connectors or to your n8n / Dify AI automation tools via custom MCP server nodes. 

When you add the server this way, you will need to authorize using OAuth with your InfraNodus API key. You can generate the key in your user account at https://infranodus.com/api-access

 

Deploying to Claude Web

  1. For Claude Web, you need to go to the Settings > Connectors menu. 
  2. Then click Add Custom Connector. 
  3. Then paste the URL above https://mcp.infranodus.com and give the server a name (e.g. InfraNodus)
  4. Click Add
  5. You will be prompted to authorize with your InfraNodus API key in the next window
  6. Once authorized, check the list of the tools available in the Connectors tab — InfraNodus should be there with the "Configure" button next to it. You can keep them with the "Needs Approval" setting for now.
  7. Open a new Claude chat and type your instruction to analyze an uploaded text, retrieve an existing graph, perform search in InfraNodus, or run a SEO analysis on a text. If Claude does not "understand" that it needs to use InfraNodus, you can explicitly ask it to use the InfraNodus tools for that. 
infranodus-mcp-claude.png

 

Deploying to n8n Cloud

In order to deploy the InfraNodus MCP server to n8n cloud, you can use the the AI Agent node and attach the MCP server to it as a tool. A custom instruction in the AI agent node will tell it to use the specific tool you provided. You can also use the official InfraNodus n8n node as well.

The basic setup looks something like this:

infranodus-n8n-workflow.png

To set it up:

  1. Create a new workflow
  2. Add a chat trigger or webhook node
  3. Add the AI Agent node, connect it to the LLM chat model you want to use it with and the Simple Memory node to keep track of conversations
  4. In the Tool connector of the AI Agent, click + and search for MCP Server
  5. Once the MCP Server node is added, you can provide the details of the InfraNodus MCP server to it.
  6. infranodus-n8n-tool-setup.png
    1. Add the URL https://mcp.infranodus.com in the Endpoint field
    2. Choose HTTP Streamable as the Server Transport method
    3. In Authentication, choose MCP OAuth2
    4. In Credential, create a new Credential
    5. You will have a new window opened where you will put the same URL https://mcp.infranodus.com as the Server URL and choose Use Dynamic Registration option above and click "Connect My Account"

    6. A new window will pop up prompting you to authorize with the InfraNodus API key for your account

      infranodus-n8n-mcp-registration.png

That's it, your n8n <> InfraNodus connection is ready. You can check our public n8n templates for some workflow examples that use the InfraNodus MCP tools.

 

Deploying to Your Local LLM App (Claude Desktop, Cursor / Windsurf IDE, local n8n / Dify)

If you're deploying the InfraNodus MCP server to your local LLM app like Claude Desktop, Claude Code, Cursor / Windsurf / VSCode IDE, or your local n8n / Dify instance, you will need to add the configurations for the server to your application's MCP config file. 

In order to do that:

  1. Find the MCP config file in your application.

  2. Add the following code to your config: 
     

    {
	"mcpServers": {
		"infranodus": {
			"command": "npx",
			"args": [
				"-y",
				"infranodus-mcp-server"
			],
  		    "env": {
			    "INFRANODUS_API_KEY": "your-api-key-here"
		    }
	    }
	}
}
  3. Note, that if you already have the mcpServers section in your config, you simply need to add the infranodus part and everything inside.
  4. In the "env" { "INFRANODUS_API_KEY" part, you need to provide your InfraNodus API key available at https://infranodus.com/api-access 
  5. Restart your app (just in case), and the MCP server is ready to use

 

What Can InfraNodus MCP Server Do?

The InfraNodus MCP, just like the web version of InfraNodus, server helps you brainstorm, refine your writing, and align it with current discourse and your audience's interests. You can:

  • Generate knowledge graphs from any text
  • Search, retrieve, and analyze your existing InfraNodus graphs
  • Analyze any text, retrieve the main topics, content gaps, and latent ideas
  • Generate research questions for any text / knowledge graph
  • Refine and develop the content of an existing text
  • Generate SEO and LLMO reports that can help you see how to embed your text into existing discourse
  • Analyze the content of search results and search intent for a query

 

InfraNodus MCP Tutorials

Here are some  tutorials that demonstrate how the InfraNodus MCP server can be used in practice:

 

 

 

 

 

Related articles

Comments

0 comments

Please sign in to leave a comment.