Arxiv Search MCP Server

🚀 arXiv Search MCP Server

An MCP (Model Context Protocol) server for searching arXiv academic papers.

This server offers a tool for searching arXiv academic papers and obtaining detailed information. It demonstrates the core concepts of MCP in the following ways:

• Representing academic paper resources using URIs and metadata.
• Providing a paper search tool with customizable parameters.
• Supporting sorting and pagination of search results.

✨ Features

Resources

• List and access papers via the arxiv://paper/{id} URI.
• Each paper includes a title, authors, abstract, publication date, and PDF link.
• Use the JSON mime type for structured content access.

Tools

• search_papers - Search for academic papers on arXiv.
  • Requires query as a mandatory parameter.
  • Optional parameters include the maximum number of results, sorting method, and pagination.
  • Returns formatted results with detailed paper information.

🔧 Development

Install dependencies:

npm install

Build the server:

npm run build

Development mode for automatic rebuilds:

npm run watch

📦 Installation

To use with the Claude desktop app, add the server configuration:

MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "arxiv-search": {
      "command": "node",
      "args": ["/path/to/arxiv-search/build/index.js"],
      "disabled": false,
      "alwaysAllow": []
    }
  }
}

To use with VSCode and Roo Cline, add to: ~/Library/Application Support/Code/User/globalStorage/rooveterinaryinc.roo-cline/settings/cline_mcp_settings.json

Debugging

Since MCP servers communicate via stdio, debugging can be challenging. We recommend using MCP Inspector, which can be used via the following package script:

npm run inspector

The Inspector will provide a URL to access the debugging tool in the browser.

💻 Usage Examples

Basic Usage

<use_mcp_tool>
<server_name>arxiv-search</server_name>
<tool_name>search_papers</tool_name>
<arguments>
{
  "query": "机器学习",
  "max_results": 5,
  "sort_by": "relevance"
}
</arguments>
</use_mcp_tool>

Available parameters:

• query (required): Search query (e.g., 'Machine Learning', 'Quantum Physics').
• max_results (optional): Maximum number of results to return (default: 10, max: 100).
• sort_by (optional): Sorting method ('relevance', 'lastUpdatedDate', 'submittedDate').
• start (optional): Starting index of results (for pagination).