Google Scholar MCP Server

<div align="center">

The most comprehensive Google Scholar MCP server for AI assistants

Installation • Quick Start • Tools • API Reference • Contributing

</div>

---

A powerful Model Context Protocol (MCP) server that connects AI assistants like Claude to Google Scholar. Search academic publications, find researchers, analyze citations, download papers, and perform advanced research workflows.

🌟 Features

Tools (11 Total)

Tool	Description
search_publications	Search for academic papers by topic, author, or date range
search_author	Find researchers by name, field, or institution
get_author_profile	Get detailed author info including h-index, publications, and coauthors
get_citations	Find papers that cite a specific publication
generate_bibtex	Generate BibTeX entries for citations
get_related_articles	Find related articles for a publication
get_all_versions	Get all available versions of a paper
download_paper	Download and store papers locally for offline access
list_papers	List all locally stored papers
read_paper	Read content of a stored paper
advanced_search	Advanced search with language, patent, and date filters

Prompts (6 Total)

Prompt	Description
literature_review	Structured approach to conducting literature reviews
find_expert	Find leading researchers in a specific field
citation_analysis	Analyze the citation impact of a publication
deep_paper_analysis	Comprehensive multi-step paper analysis workflow
research_synthesis	Synthesize research across multiple papers
methodology_comparison	Compare methodologies across research papers

Key Advantages

• 11 Powerful Tools: Search, citations, related articles, versions, downloads & more
• Advanced Search: Filter by language, patents, review articles, and more
• Local Paper Storage: Download and cache papers for offline access
• Intelligent Caching: Reduce redundant requests with TTL-based caching
• Rich Metadata: Abstracts, citation counts, h-index, coauthors, and more
• Rate Limiting: Built-in protection against Google Scholar blocking
• Error Handling: Graceful handling of rate limits and CAPTCHAs
• BibTeX Support: Generate proper academic citations
• Advanced Prompts: Pre-built workflows for deep paper analysis and research synthesis

📦 Installation

Via Smithery (Recommended)

npx -y @smithery/cli install @dihannahdi/google-scholar-mcp --client claude

Via npm

npm install -g @dihannahdi/google-scholar-mcp

From Source

git clone https://github.com/dihannahdi/google-scholar-mcp.git
cd google-scholar-mcp
npm install
npm run build

🚀 Quick Start

Running the Server

# If installed globally
google-scholar-mcp

# From source
npm start

# Development mode
npm run dev

Configuration for Claude Desktop

Add to your claude_desktop_config.json:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "google-scholar": {
      "command": "node",
      "args": ["/path/to/google-scholar-mcp/dist/index.js"],
      "env": {
        "SCHOLAR_STORAGE_PATH": "~/.google-scholar-mcp/papers",
        "SCHOLAR_RATE_LIMIT_MS": "3000",
        "SCHOLAR_CACHE_ENABLED": "true",
        "SCHOLAR_CACHE_TTL_MS": "3600000",
        "SERPAPI_KEY": "your_serpapi_key_here",
        "SCHOLAR_USE_SERPAPI_FALLBACK": "true"
      }
    }
  }
}

Or using npx:

{
  "mcpServers": {
    "google-scholar": {
      "command": "npx",
      "args": ["-y", "@dihannahdi/google-scholar-mcp"]
    }
  }
}

Environment Variables

Variable	Description	Default
SCHOLAR_STORAGE_PATH	Path to store downloaded papers	~/.google-scholar-mcp/papers
SCHOLAR_RATE_LIMIT_MS	Rate limit between requests (ms)	3000
SCHOLAR_CACHE_ENABLED	Enable response caching	true
SCHOLAR_CACHE_TTL_MS	Cache time-to-live (ms)	3600000 (1 hour)
SCHOLAR_PROXY_URL	Optional proxy URL	-
SERPAPI_KEY	SerpAPI key for reliable fallback (get free key)	-
SCHOLAR_USE_SERPAPI_FALLBACK	Enable automatic SerpAPI fallback when blocked	false
SCHOLAR_ROTATE_USER_AGENT	Rotate user agents to avoid detection	true
SCHOLAR_ENABLE_JITTER	Add random delays between requests	true
SCHOLAR_JITTER_MAX_MS	Maximum jitter delay in milliseconds	2000

🔐 SerpAPI Fallback (Recommended)

Google Scholar aggressively blocks automated access. To ensure 100% uptime, we recommend using SerpAPI as a fallback:

1. Get a free API key at serpapi.com (100 free searches/month)
2. Set environment variables:

   export SERPAPI_KEY="your_serpapi_key"
   export SCHOLAR_USE_SERPAPI_FALLBACK="true"
   

How it works:

• The server first attempts direct scraping (free, but may be blocked)
• If blocked, it automatically falls back to SerpAPI (reliable, paid)
• Results are normalized to the same format regardless of source

Benefits of SerpAPI:

• ✅ 100% uptime guarantee
• ✅ No CAPTCHA or blocking issues
• ✅ Handles all anti-bot measures
• ✅ Rich structured data
• ✅ Author profiles, citations, related articles

📖 Usage Examples

Search for Publications

Search for recent machine learning papers about transformers

Tool: search_publications
Arguments:
  query: "transformer neural network"
  yearStart: 2020
  numResults: 10
  sortBy: "relevance"

Find an Author

Find researchers working on quantum computing at MIT

Tool: search_author
Arguments:
  query: "quantum computing"
  organization: "MIT"
  numResults: 5

Get Author Profile

Get detailed profile for Geoffrey Hinton

Tool: get_author_profile
Arguments:
  scholarId: "JicYPdAAAAAJ"

Get Citations

Find papers citing "Attention Is All You Need"

Tool: get_citations
Arguments:
  clusterId: "4054916225996727837"
  numResults: 10
  sortBy: "date"

Generate BibTeX

Create BibTeX for a paper

Tool: generate_bibtex
Arguments:
  title: "Attention Is All You Need"
  authors: ["Ashish Vaswani", "Noam Shazeer", "Niki Parmar"]
  year: 2017
  venue: "Advances in Neural Information Processing Systems"

🔧 API Reference

search_publications

Search Google Scholar for academic publications.

Parameters:

Parameter	Type	Required	Default	Description
query	string	Yes	-	Search query
author	string	No	-	Filter by author name
yearStart	number	No	-	Start year for filter
yearEnd	number	No	-	End year for filter
numResults	number	No	10	Number of results (1-20)
sortBy	string	No	"relevance"	"relevance" or "date"

Returns:

• List of publications with title, authors, abstract, venue, year, citation count, URLs

search_author

Search for academic researchers.

Parameters:

Parameter	Type	Required	Default	Description
query	string	Yes	-	Author name or keywords
organization	string	No	-	Filter by institution
numResults	number	No	10	Number of results (1-20)

Returns:

• List of author profiles with name, affiliation, interests, citation count

get_author_profile

Get detailed author information.

Parameters:

Parameter	Type	Required	Description
scholarId	string	Yes	Google Scholar author ID

Returns:

• Complete profile with publications, h-index, i10-index, coauthors, citation history

get_citations

Get papers that cite a specific publication.

Parameters:

Parameter	Type	Required	Default	Description
clusterId	string	Yes	-	Google Scholar cluster ID
numResults	number	No	10	Number of results (1-20)
sortBy	string	No	"relevance"	"relevance" or "date"

Returns:

• List of citing publications with details

generate_bibtex

Generate a BibTeX citation entry.

Parameters:

Parameter	Type	Required	Description
title	string	Yes	Publication title
authors	string[]	Yes	List of author names
year	number	No	Publication year
venue	string	No	Journal/conference name
url	string	No	Publication URL

Returns:

• Formatted BibTeX entry

get_related_articles

Find related articles for a given publication.

Parameters:

Parameter	Type	Required	Default	Description
clusterId	string	Yes	-	Google Scholar cluster ID
numResults	number	No	10	Number of results (1-20)

Returns:

• List of related publications with metadata

get_all_versions

Get all available versions of a paper.

Parameters:

Parameter	Type	Required	Default	Description
clusterId	string	Yes	-	Google Scholar cluster ID
numResults	number	No	10	Number of results (1-20)

Returns:

• List of all versions (preprint, published, etc.)

download_paper

Download and store a paper locally.

Parameters:

Parameter	Type	Required	Description
url	string	Yes	URL to the paper (PDF, arXiv, etc.)
filename	string	Yes	Name for the saved file
metadata	object	No	Optional metadata (title, authors, year)

Returns:

• Path to stored file and confirmation

list_papers

List all locally stored papers.

Parameters:

Parameter	Type	Required	Description
pattern	string	No	Optional filter pattern

Returns:

• List of stored papers with metadata

read_paper

Read the content of a stored paper.

Parameters:

Parameter	Type	Required	Description
filename	string	Yes	Name of the stored file

Returns:

• Paper content (text format)

advanced_search

Advanced search with additional filters.

Parameters:

Parameter	Type	Required	Default	Description
query	string	Yes	-	Search query
exactPhrase	string	No	-	Exact phrase to match
withoutWords	string	No	-	Words to exclude
author	string	No	-	Filter by author
source	string	No	-	Filter by source/journal
yearStart	number	No	-	Start year filter
yearEnd	number	No	-	End year filter
language	string	No	"en"	Language code
includePatents	boolean	No	true	Include patents in results
includeCitations	boolean	No	true	Include citations
numResults	number	No	10	Number of results

Returns:

• List of publications matching criteria

🛠️ Development

Project Structure

google-scholar-mcp/
├── src/
│   ├── index.ts          # MCP server entry point
│   ├── config/
│   │   └── index.ts      # Environment configuration
│   ├── types/
│   │   └── index.ts      # TypeScript type definitions
│   ├── tools/
│   │   ├── definitions.ts # Tool schemas (11 tools)
│   │   ├── handlers.ts    # Tool implementation
│   │   └── index.ts
│   ├── scraper/
│   │   ├── scholar.ts    # Google Scholar scraper
│   │   └── index.ts
│   └── utils/
│       ├── helpers.ts    # URL builders, rate limiting
│       ├── cache.ts      # In-memory caching
│       ├── storage.ts    # Local paper storage
│       └── index.ts
├── dist/                  # Compiled JavaScript
├── package.json
├── tsconfig.json
└── README.md

Building

npm run build

Testing

# Run MCP Inspector for debugging
npx @modelcontextprotocol/inspector node dist/index.js

⚠️ Limitations

1. Rate Limiting: Google Scholar may rate-limit or block requests. The server includes delays and retries, but heavy usage may trigger blocks.

2. CAPTCHA: Excessive requests may trigger CAPTCHA challenges. If this happens, wait a few minutes before retrying.

3. No Official API: This server scrapes Google Scholar's web interface, which may break if Google changes their HTML structure.

4. Results Limited: Maximum 20 results per request to stay within reasonable limits.

🤝 Contributing

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

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Model Context Protocol for the MCP SDK
• Scholarly for inspiration
• The academic community for making research accessible

📬 Support

For issues and questions:

• Open an issue on GitHub
• Check the MCP documentation

---

Made with ❤️ for the research community