papers-mcp by guangxiangdebizi - MCP Server

📚 Paper MCP

A Model Context Protocol (MCP) server for academic paper search and analysis. This tool helps researchers quickly find, analyze, and extract information from academic papers using open data sources.

✨ Features

🔍 Paper Search (`papers.search`)

Multi-source search: OpenAlex, Crossref, arXiv
Advanced filtering: Author, affiliation, venue, year range, field of study
Smart deduplication: Automatically merges results from different sources
Open access focus: Option to filter for freely available papers
Flexible sorting: By relevance, date, or citation count

📖 Paper Analysis (`papers.analyze`)

Universal URL support: PDF links, arXiv pages, DOI links, journal pages
Intelligent extraction: Title, authors, abstract, metadata
Full-text analysis: Optional complete content extraction
Multiple formats: Handles PDF, HTML, and XML sources
Rich metadata: Publication details, citation info, access links

🚀 Quick Start

Installation

# Clone the repository
git clone https://github.com/guangxiangdebizi/paper-mcp.git
cd paper-mcp

# Install dependencies
npm install

# Build the project
npm run build

Usage

Local Development (Stdio)

# Start the MCP server
npm start

Web Interface (SSE)

# Install supergateway globally
npm install -g supergateway

# Start SSE server on port 3100
npm run sse

Claude Desktop Configuration

Stdio Mode

Add to your Claude Desktop config:

{
  "mcpServers": {
    "paper-mcp": {
      "command": "node",
      "args": ["path/to/paper-mcp/build/index.js"]
    }
  }
}

SSE Mode

{
  "mcpServers": {
    "paper-mcp": {
      "type": "sse",
      "url": "http://localhost:3100/sse",
      "timeout": 600
    }
  }
}

🛠️ API Reference

Search Papers

// Basic search
{
  "query": "machine learning transformers"
}

// Advanced search with filters
{
  "query": "neural networks",
  "authors": ["Geoffrey Hinton", "Yann LeCun"],
  "affiliations": ["Google", "Meta"],
  "venue": ["Nature", "Science"],
  "year_from": 2020,
  "year_to": 2024,
  "open_access_only": true,
  "sort_by": "citations_desc",
  "page": 1,
  "page_size": 20
}

Analyze Paper

// Analyze from URL
{
  "url": "https://arxiv.org/abs/2301.00001"
}

// Full-text extraction
{
  "url": "https://example.com/paper.pdf",
  "include_full_text": true,
  "max_content_length": 10000
}

🌐 Data Sources

OpenAlex: Comprehensive academic database
Crossref: DOI and metadata registry
arXiv: Preprint repository
Unpaywall: Open access link detection

📊 Example Outputs

Search Results

{
  "page": 1,
  "page_size": 20,
  "has_more": true,
  "items": [
    {
      "title": "Attention Is All You Need",
      "authors": ["Ashish Vaswani", "Noam Shazeer"],
      "venue": "NeurIPS",
      "year": 2017,
      "doi": "10.5555/3295222.3295349",
      "primary_url": "https://arxiv.org/abs/1706.03762",
      "oa_pdf_url": "https://arxiv.org/pdf/1706.03762.pdf",
      "citation_count": 50000,
      "source": "openalex"
    }
  ]
}

Analysis Results

{
  "title": "Attention Is All You Need",
  "authors": ["Ashish Vaswani", "Noam Shazeer"],
  "abstract": "The dominant sequence transduction models...",
  "content": "Full paper content...",
  "type": "arxiv",
  "metadata": {
    "arxiv_id": "1706.03762",
    "published": "2017-06-12",
    "categories": ["cs.CL", "cs.AI"]
  }
}

🔧 Development

# Watch mode for development
npm run dev

# Build for production
npm run build

# Run tests (if available)
npm test

📝 License

Apache 2.0 - see file for details.

👨‍💻 Author

Xingyu Chen

📧 Email: guangxiangdebizi@gmail.com
🔗 LinkedIn: xingyu-chen-b5b3b0313
🐙 GitHub: @guangxiangdebizi
📦 NPM: @xingyuchen

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

🙏 Acknowledgments

OpenAlex for providing comprehensive academic data
Crossref for DOI resolution services
arXiv for open access to preprints
The MCP community for the excellent protocol

Built with ❤️ for the research community