PythAIs

You run a Node.js-based MCP server that connects to the Pythagraph RED API to fetch, analyze, and present graph data. This server automates data retrieval, computes statistics, formats outputs into readable tables, and provides concise summaries to help you quickly understand graph structures and relationships.

How to use

To use this MCP server, you run it as an MCP endpoint and connect to it with your MCP client. You can fetch detailed graph data for a specific graphId, or request a quick summary of the graph. The server formats results into easy-to-read tables and highlights key insights such as node/edge type distributions and density.

Common workflows include requesting full graph data to inspect nodes, edges, and metadata, or obtaining a concise overview to compare multiple graphs at a glance. The server handles errors such as invalid graph IDs and API timeouts, and presents the outputs in a consistent, machine-friendly format for downstream processing.

How to install

Prerequisites: you need Node.js installed on your system. You can verify Node.js and npm are available by running the following commands.

Check Node.js and npm versions to ensure compatibility:

node -v
npm -v

Install and run via NPX

Use NPX to run the MCP server without installing it globally. This is convenient for quick experiments or evaluation.

npx -y pythais-mcp-server

Install and run via Docker

Run the MCP server in a Docker container for isolation and portability.

docker run -i --rm mcp/pythais

Alternate local setup (build steps)

If you prefer a local build/run flow, you can install dependencies, build, and start the server using npm scripts. This option is useful for development and debugging.

npm install
npm run build
npm test

Configuring how to connect to the Pythagraph RED API

The server targets the Pythagraph RED API at the following endpoint. You supply the graphId when requesting data. Use the endpoint format shown here when configuring your client requests.

Endpoint used by the server (example):

https://red.pythagraph.co.kr/api/red/graph/exportGraphInfo.do?graphId={graphId}

Notes on usage and outputs

The server provides two main outputs you can request: a detailed graph data view and a concise graph summary. The detailed view includes tables with node/edge distributions and the first 10 nodes and edges. The summary view focuses on key insights like the highest/lowest values, total counts, and a quick density estimate.

Error handling is built in to manage invalid IDs, API timeouts, and unexpected responses. Outputs are formatted for readability and downstream processing.

Security and maintenance

Keep dependencies up to date and monitor API access to avoid timeouts or unauthorized usage. If you expose the MCP endpoint publicly, consider implementing appropriate authentication and access controls.

Troubleshooting

If you encounter timeouts or invalid responses, check that the graphId you provide exists in the Pythagraph RED system and that the API URL is reachable from your environment. Review error messages for hints about authentication, network issues, or malformed requests.

Available tools

get_graph_data

Retrieve detailed graph data from the Pythagraph RED API, including nodes, edges, statistics, and metadata formatted as tables and descriptions.

get_graph_summary

Get a concise overview of the graph, including counts, distributions, and key insights; supports an optional detailed view.