arXiv MCP ServerServer

prashalruchiranga
GitHub
ResearchData RetrievalNatural Language Processing

Loading subscription status...

💡 Description

A Model Context Protocol (MCP) server that allows interaction with arXiv.org using natural language. It enables users to search, retrieve metadata and download scholarly articles in PDF format, and load articles into large language models (LLMs) for analysis.

📝 JSON Entries

{
  "mcpServers": [
    {
      "arxiv-server": {
        "env": {
          "DOWNLOAD_PATH": "/ABSOLUTE/PATH/TO/DOWNLOADS/FOLDER"
        },
        "args": [
          "--directory",
          "/ABSOLUTE/PATH/TO/PARENT/FOLDER/arxiv-mcp-server/src/arxiv_server",
          "run",
          "server.py"
        ],
        "command": "uv"
      }
    }
  ]
}

🛠️ Tools

Get article URLDownload articleLoad article into contextGet detailsSearch arXiv

Features

  • Retrieve metadata about scholarly articles
  • Download articles as PDFs
  • Search arXiv database with specific queries
  • Load articles into LLM context for analysis

💬 Example Queries

  • Can you get the details of 'Reasoning to Learn from Latent Thoughts' paper?
  • Get the papers authored or co-authored by Yann Lecun on convolutional neural networks
  • Download the attention is all you need paper
  • Can you get the papers by Andrew NG which have 'convolutional neural networks' in title?
  • Can you display the paper?
  • List the titles of papers by Yann LeCun. Paginate through the API until there are 30 titles