Skip to main content

arXiv 논문 데이터를 Claude AI와 연동하는 MCP 서버

Project description

🧠 arXiv Research Assistant MCP Server

smithery badge

This project is an MCP (Model Context Protocol) server built to interact with the vast arXiv.org paper database.

It allows clients like Claude AI to search, explore, and compare arXiv papers efficiently — all through a custom-built, local server. It’s built with Python and the FastMCP framework, and uses uv for lightweight package management.

arXiv Research Assistant Server MCP server MseeP.ai Security Assessment Badge

✨ Features

  • 🔍 Keyword-based Paper Search
    Search arXiv papers by keywords, with options to sort by relevance or most recent.

  • 📚 Latest Papers by Category
    Specify an arXiv category code (e.g., cs.AI, math.AP) to fetch the most recent papers in that field.

  • 📄 Paper Details Lookup
    Fetch detailed metadata using a paper's arXiv ID: title, authors, abstract, categories, DOI, PDF link, and more.

  • 🧑‍🔬 Author-based Paper Search
    Retrieve a list of papers published by a specific author.

  • 📊 Trend Analysis (Experimental)
    Get an overview of trending keywords or topics based on recent papers in a category (currently uses mock data).

  • 📝 Summarization Prompt Generator
    Dynamically generate prompts that help LLMs summarize a selected paper more effectively.

  • 🆚 Comparison Prompt Generator
    Provide two paper IDs to generate a structured prompt for comparing their content.


🛠️ Tech Stack

  • Python 3.11+
  • FastMCP
  • uv (for dependency & environment management)
  • requests (for API communication)
  • xml.etree.ElementTree (for parsing XML responses)

🚀 Getting Started

Installing via Smithery

To install arXiv Research Assistant MCP Server for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install arxiv-paper-mcp --client claude

Installation from PyPI

uv pip install arxiv-paper-mcp

🔧 Clone the repository (for development)

git clone https://github.com/daheepk/arxiv-mcp-server.git
cd arxiv-mcp-server

🔧 Install Dependencies (for development)

Use uv to install all dependencies in editable mode:

uv pip install -e .

⚙️ How to Run

▶️ Run the server (locally)

arxiv-paper-mcp

🔌 Use with Claude

To use this MCP server with Claude, add the following JSON configuration to Claude's MCP settings:

{
  "mcpServers": {
    "arXivPaper": {
      "command": "uv",
      "args": [
        "tool",
        "run",
        "arxiv-paper-mcp"
      ]
    }
  }
}

Project Structure

arxiv-mcp-server/
├── arxiv_mcp/              # Main package
│   ├── __init__.py
│   ├── app.py              # FastMCP app setup
│   ├── server.py           # Server entry point
│   ├── utils.py            # arXiv API communication logic
│   ├── resources/          # MCP resources (categories, authors, etc.)
│   ├── tools/              # MCP tools (search, detail lookup, trends)
│   └── prompts/            # Prompt templates (summarize, compare)
├── pyproject.toml          # Project config & dependencies
└── README.md               # This file

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

iflow_mcp_daheepk_arxiv_paper_mcp-0.1.3.tar.gz (11.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

File details

Details for the file iflow_mcp_daheepk_arxiv_paper_mcp-0.1.3.tar.gz.

File metadata

File hashes

Hashes for iflow_mcp_daheepk_arxiv_paper_mcp-0.1.3.tar.gz
Algorithm Hash digest
SHA256 1e1ca81c74f7bdf02c4fe70291d99c3d7205f37ebc6817a999ce5ab6058e7c62
MD5 d3cf3ece06e24af51ed3c79f33bb8b93
BLAKE2b-256 42a7c64eb9fb640e38fb76208375155e3db8a44cd6827cc308f68597548edbfe

See more details on using hashes here.

File details

Details for the file iflow_mcp_daheepk_arxiv_paper_mcp-0.1.3-py3-none-any.whl.

File metadata

File hashes

Hashes for iflow_mcp_daheepk_arxiv_paper_mcp-0.1.3-py3-none-any.whl
Algorithm Hash digest
SHA256 eb4b502b53eadafd5a011f9d5696e32d79d72e12f3c7cc63703618e918178303
MD5 b0a74ac72798cea5d4ab8f087c65e700
BLAKE2b-256 06b3047ead811e75853dcfca05368749e4236a2d64a35ef5720ec5dc13c26d7f

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page