arXiv Search MCP Server

What is arXiv Search MCP Server?

A Model Context Protocol (MCP) server that enables natural language searching of arXiv academic papers. Built with Node.js and TypeScript for local execution, it works with Claude Desktop and other MCP-compatible clients to fetch live data from the arXiv API.

How to use arXiv Search MCP Server?

Clone the repository, run npm install and npm run build, then add the server to your Claude Desktop configuration by pointing to the built dist/index.js file. After setup, use natural language phrases such as “最新の深層学習研究を5件見つけて” or “Geoffrey Hintonの機械学習論文” to search papers.

Key features of arXiv Search MCP Server

• Natural language queries in English and Japanese
• Search by keywords, authors, categories, and date ranges
• Structured results with titles, authors, abstracts, and links
• MCP protocol compliance for seamless client integration
• Robust error handling and retry logic for API calls
• Real-time data fetched directly from the arXiv API

Use cases of arXiv Search MCP Server

• Find the latest deep learning research papers using conversational language
• Locate papers by a specific author, e.g., Geoffrey Hinton on machine learning
• Search for recent studies on a topic like quantum computing since 2023
• Retrieve detailed information about a known paper using its arXiv ID