Skip to main content
Glama

Academic MCP Server

by nanyang12138
GitHub
Python
  • Apple
  • Linux
OverviewInspectNewEndpointsSchemaRelated ServersReviewsScore
Need Help?View Source CodeReport Issue

Academic MCP Server

šŸ” A unified Model Context Protocol (MCP) server that provides AI assistants access to multiple academic databases through a single, consistent interface.

šŸŒŸ Features

Supported Databases

  • PubMed šŸ„ - Biomedical and life sciences literature (NCBI)

  • bioRxiv šŸ§¬ - Biology preprints

  • medRxiv šŸ’Š - Medical preprints

  • arXiv šŸ”¬ - Physics, mathematics, computer science, and more

  • Semantic Scholar šŸ¤– - AI-powered academic search across disciplines

Core Capabilities

  • āœ… Unified Search: Search across all databases with a single query

  • āœ… Advanced Filtering: Filter by title, author, date, journal, and more

  • āœ… Metadata Access: Retrieve detailed paper information

  • āœ… PDF Download: Download open access papers when available

  • āœ… Deep Analysis: Generate comprehensive paper analysis prompts

  • āœ… Standardized Output: Consistent data format across all sources

šŸš€ Quick Start

Prerequisites

  • Python 3.10+

  • FastMCP library

  • Internet connection

Installation

āœ… Already Installed! Your Academic MCP Server is fully configured and ready to use.

If you need to set it up on another machine:

  1. Clone or download this repository:

    cd Academic-MCP-Server

  2. Create a virtual environment:

    python -m venv venv

  3. Activate the virtual environment:

    • Windows: venv\Scripts\activate

    • Mac/Linux: source venv/bin/activate

  4. Install dependencies:

    pip install -r requirements.txt

Note: All PubMed functionality is integrated locally. No external dependencies required!

Configuration for Cursor

This project provides TWO MCP servers with complementary features:

  1. academic - Basic search, metadata retrieval, and PDF downloads across 5 databases

  2. academic-research - Advanced features including citation analysis, paper impact evaluation, local PDF analysis, and complete research workflows

Add this configuration to your MCP settings file (~/.cursor/mcp.json or C:\Users\YOUR_USERNAME\.cursor\mcp.json):

Windows:

{ "mcpServers": { "academic": { "command": "C:\\Users\\YOUR_USERNAME\\path\\to\\Academic-MCP-Server\\venv\\Scripts\\python.exe", "args": [ "C:\\Users\\YOUR_USERNAME\\path\\to\\Academic-MCP-Server\\academic_server.py" ], "env": {}, "disabled": false, "autoApprove": [] }, "academic-research": { "command": "C:\\Users\\YOUR_USERNAME\\path\\to\\Academic-MCP-Server\\venv\\Scripts\\python.exe", "args": [ "C:\\Users\\YOUR_USERNAME\\path\\to\\Academic-MCP-Server\\academic_research_advanced.py" ], "env": {}, "disabled": false, "autoApprove": [] } } }

Mac/Linux:

{ "mcpServers": { "academic": { "command": "/path/to/Academic-MCP-Server/venv/bin/python", "args": [ "/path/to/Academic-MCP-Server/academic_server.py" ], "env": {}, "disabled": false, "autoApprove": [] }, "academic-research": { "command": "/path/to/Academic-MCP-Server/venv/bin/python", "args": [ "/path/to/Academic-MCP-Server/academic_research_advanced.py" ], "env": {}, "disabled": false, "autoApprove": [] } } }

Note: Replace YOUR_USERNAME and path/to with your actual paths.

šŸ“– Usage

Search Papers

Search across all databases:

search_papers( keywords="UCAR-T", source="all", num_results=15 )

Search specific database:

search_papers( keywords="machine learning", source="arxiv", num_results=10 )

Advanced Search

search_papers_advanced( title="neural networks", author="Hinton", start_date="2020-01-01", end_date="2024-12-31", source="semantic_scholar", num_results=10 )

PubMed-specific advanced search:

search_papers_advanced( title="CAR-T", author="Wang", journal="Nature", start_date="2024/01/01", # PubMed uses YYYY/MM/DD end_date="2025/12/31", source="pubmed", num_results=10 )

Get Paper Metadata

# PubMed get_paper_metadata(identifier="40883768", source="pubmed") # bioRxiv get_paper_metadata(identifier="10.1101/2024.01.001", source="biorxiv") # arXiv get_paper_metadata(identifier="2301.00001", source="arxiv") # Semantic Scholar (Paper ID or DOI) get_paper_metadata(identifier="DOI:10.1038/s41586-020-1234-5", source="semantic_scholar")

Download PDF

download_paper_pdf(identifier="2301.00001", source="arxiv")

List Available Sources

list_available_sources() # Returns: ["pubmed", "biorxiv", "medrxiv", "arxiv", "semantic_scholar"]

Deep Paper Analysis

deep_paper_analysis(identifier="40883768", source="pubmed")

šŸ›  MCP Tools Reference

Server: academic (Basic Search & Retrieval)

1. search_papers

Search for papers using keywords.

Parameters:

  • keywords (str): Search query

  • source (str): "all", "pubmed", "biorxiv", "medrxiv", "arxiv", or "semantic_scholar"

  • num_results (int): Number of results per source (default: 10)

2. search_papers_advanced

Advanced search with multiple filters.

Parameters:

  • title (str, optional): Search in titles

  • author (str, optional): Author name

  • journal (str, optional): Journal name

  • start_date (str, optional): Start date

  • end_date (str, optional): End date

  • term (str, optional): General search term

  • source (str): Database source

  • num_results (int): Number of results

3. get_paper_metadata

Get detailed metadata for a specific paper.

Parameters:

  • identifier (str): Paper ID (PMID, DOI, arXiv ID, etc.)

  • source (str): Database source

4. download_paper_pdf

Download PDF for a paper.

Parameters:

  • identifier (str): Paper ID

  • source (str): Database source

5. list_available_sources

List all available databases.

6. deep_paper_analysis

Generate comprehensive analysis prompt.

Parameters:

  • identifier (str): Paper ID

  • source (str): Database source

Server: academic-research (Advanced Analysis & Research)

1. analyze_citation_network

Analyze paper's citation network.

Parameters:

  • paper_id (str): Paper identifier (DOI, PMID, etc.)

  • source (str): Data source (default: "semantic_scholar")

  • max_depth (int): Network depth 1-3 layers (default: 2)

2. evaluate_paper_impact

Evaluate academic impact of a paper.

Parameters:

  • paper_id (str): Paper identifier

  • source (str): Data source (default: "semantic_scholar")

3. recommend_related_papers

Recommend related papers using multiple strategies.

Parameters:

  • paper_id (str): Source paper identifier

  • source (str): Data source (default: "semantic_scholar")

  • num_recommendations (int): Number of recommendations (default: 10)

  • strategy (str): "comprehensive", "citations", "similar", or "influential"

4. research_workflow_complete

ā­ Recommended Core Feature - Complete research workflow: retrieve ā†’ analyze ā†’ read ā†’ summarize

Parameters:

  • topic (str): Research topic (e.g., "CRISPR gene editing")

  • num_papers (int): Number of papers to retrieve (default: 5)

  • include_analysis (bool): Include deep analysis (default: true)

  • include_summary (bool): Include auto-summary (default: true)

5. analyze_local_paper

Comprehensively analyze local or online PDF papers.

Parameters:

  • pdf_path (str): PDF file path (local or URL)

  • include_figures (bool): Analyze figures (default: true)

  • include_summary (bool): Generate summary (default: true)

6. list_all_figures

List all figures from a PDF paper.

Parameters:

  • pdf_path (str): PDF file path (local or URL)

7. explain_specific_figure

Explain a specific figure from a PDF.

Parameters:

  • pdf_path (str): PDF file path (local or URL)

  • figure_number (int): Figure number (e.g., 1, 2, 3)

  • provide_context (bool): Include context paragraphs (default: true)

šŸ“Š Standardized Output Format

All search results return papers in this standardized format:

{ "id": "Unique identifier (PMID, DOI, arXiv ID, etc.)", "title": "Paper title", "authors": "Author names (comma-separated)", "abstract": "Paper abstract", "publication_date": "Publication date", "journal": "Journal or venue name", "url": "Link to paper", "pdf_url": "PDF link (if available)", "source": "Database source (pubmed/biorxiv/arxiv/etc.)" }

Semantic Scholar results include additional fields:

  • citation_count: Number of citations

  • reference_count: Number of references

  • fields_of_study: Research areas

šŸ”§ Architecture

Adapter Pattern

Each database is wrapped in an adapter that implements a common interface:

Academic-MCP-Server/ ā”œā”€ā”€ academic_server.py # Main MCP server ā”œā”€ā”€ adapters/ # Database adapters ā”‚ ā”œā”€ā”€ base_adapter.py # Abstract base class ā”‚ ā”œā”€ā”€ pubmed_adapter.py # PubMed wrapper ā”‚ ā”œā”€ā”€ biorxiv_adapter.py # bioRxiv/medRxiv ā”‚ ā”œā”€ā”€ arxiv_adapter.py # arXiv ā”‚ ā””ā”€ā”€ semantic_scholar_adapter.py ā”œā”€ā”€ utils/ # Helper functions ā””ā”€ā”€ requirements.txt

Adding New Databases

To add a new database:

  1. Create a new adapter in adapters/

  2. Inherit from BaseAdapter

  3. Implement all required methods

  4. Register in academic_server.py

Example:

# adapters/new_database_adapter.py from .base_adapter import BaseAdapter class NewDatabaseAdapter(BaseAdapter): def search_by_keywords(self, keywords, num_results): # Implementation pass # ... implement other methods # In academic_server.py from adapters.new_database_adapter import NewDatabaseAdapter adapters = { # ... existing adapters "new_database": NewDatabaseAdapter() }

šŸŽÆ Use Cases

For Researchers

  • Search across multiple preprint servers simultaneously

  • Find papers by specific authors or topics

  • Download open access papers automatically

  • Generate literature review materials

For AI Assistants

  • Access comprehensive academic knowledge

  • Provide up-to-date research information

  • Help with citation and reference management

  • Analyze research trends and findings

āš ļø Limitations & Notes

API Rate Limits

  • PubMed: No API key required, but rate-limited

  • bioRxiv/medRxiv: No authentication required

  • arXiv: Rate-limited (1 request per 3 seconds recommended)

  • Semantic Scholar: Free tier has rate limits; get API key for higher limits at https://www.semanticscholar.org/product/api

PDF Availability

  • PubMed: Only PMC open access articles

  • bioRxiv/medRxiv: All articles are open access

  • arXiv: All articles are open access

  • Semantic Scholar: Depends on publisher policies

Date Formats

  • PubMed: YYYY/MM/DD

  • Others: YYYY-MM-DD

šŸ¤ Contributing

Contributions are welcome! Feel free to:

  • Add new database adapters

  • Improve existing functionality

  • Fix bugs

  • Enhance documentation

šŸ“„ License

This project builds upon the PubMed-MCP-Server and follows similar open-source principles.

šŸ™ Acknowledgments

  • PubMed-MCP-Server for the original PubMed integration

  • NCBI E-utilities

  • bioRxiv/medRxiv API

  • arXiv API

  • Semantic Scholar API

  • FastMCP framework

šŸ“ž Support

For issues or questions:

  1. Check the documentation above

  2. Review error messages in logs

  3. Ensure all dependencies are installed

  4. Verify your MCP configuration

Happy researching! šŸ“ššŸ”¬

This server cannot be installed

-
security - not tested
F
license - not found
-
quality - not tested

How are these scores calculated?

hybrid server

The server is able to function both locally and remotely, depending on the configuration or use case.

Enables AI assistants to search across multiple academic databases (PubMed, arXiv, bioRxiv, medRxiv, Semantic Scholar) through a unified interface. Supports advanced filtering, metadata retrieval, PDF downloads, and comprehensive research workflows with citation analysis.

  1. šŸŒŸ Features
    1. Supported Databases
    2. Core Capabilities
  2. šŸš€ Quick Start
    1. Prerequisites
    2. Installation
    3. Configuration for Cursor
  3. šŸ“– Usage
    1. Search Papers
    2. Advanced Search
    3. Get Paper Metadata
    4. Download PDF
    5. List Available Sources
    6. Deep Paper Analysis
  4. šŸ›  MCP Tools Reference
    1. Server: academic (Basic Search & Retrieval)
    2. Server: academic-research (Advanced Analysis & Research)
  5. šŸ“Š Standardized Output Format
    1. šŸ”§ Architecture
      1. Adapter Pattern
      2. Adding New Databases
    2. šŸŽÆ Use Cases
      1. For Researchers
      2. For AI Assistants
    3. āš ļø Limitations & Notes
      1. API Rate Limits
      2. PDF Availability
      3. Date Formats
    4. šŸ¤ Contributing
      1. šŸ“„ License
        1. šŸ™ Acknowledgments
          1. šŸ“ž Support

            New MCP Servers

            MCP directory API

            We provide all the information about MCP servers via our MCP API.

            curl -X GET 'https://glama.ai/api/mcp/v1/servers/nanyang12138/Academic-MCP-Server'

            If you have feedback or need assistance with the MCP directory API, please join our Discord server