🏛️ ArXiv MCP: A Humanity Offering

Empowering the global research community through high-performance, AI-native scientific discovery.

🌎 The Vision

This repository is a Humanity Offering—a tools-first contribution designed to lower the friction between human curiosity and the vast expanse of scientific knowledge. By bridging the Model Context Protocol (MCP) with ArXiv, we enable AI assistants to not just "search" for papers, but to deeply read, analyze, and synthesize the latest breakthroughs in real-time.

✨ Elite Features

• 🔍 Ultra-Precision Search: Advanced filtering and sorting across 2M+ papers.
• 📄 Deep Content Extraction: Automated PDF-to-Markdown conversion. Unlike other tools, this lets AI models "read" the full text, figures, and methodology of any paper.
• ⚡ Proactive Caching: High-performance local caching for instant retrieval of previously analyzed papers.
• ✍️ Submission Catalyst: Programmatic helper tools for ArXiv's SWORD API, facilitating the path from reading research to contributing it.
• 🧩 MCP Native: Seamless integration with modern AI environments (Claude Desktop, Antigravity, etc.).

🚀 Deployment

Direct execution (via uv)

uvx arxiv-mcp

Manual Installation

git clone https://github.com/your-username/arxiv-mcp.git
cd arxiv-mcp
pip install .

🛠️ Toolset

| Tool | Capability | |------|------------| | search_papers | Dynamic search with relevance and date sorting. | | get_paper_content | The Core Utility: Full PDF extraction to clean Markdown. | | list_recent_papers | Real-time monitoring of specific ArXiv categories. | | generate_submission_metadata | Generates Atom XML for programmatic ArXiv submissions. | | get_details | Comprehensive metadata and abstract retrieval. |

🕹️ Configuration

Claude Desktop

Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "arxiv-humanity": {
      "command": "python3",
      "args": ["-m", "src.arxiv_mcp.server"],
      "cwd": "/absolute/path/to/arxiv-mcp",
      "env": {
        "PYTHONPATH": "src"
      }
    }
  }
}

📜 Ethical Commitment

This project is released under the MIT License. It is intended for the ethical advancement of knowledge. We encourage the community to build upon this bridge, improve the extraction algorithms, and share findings openly.