Overview
Search and retrieve academic papers from arXiv with full-text access.
Setup
Run with npx:
npx -y @modelcontextprotocol/server-arxivConfiguration
No configuration requiredDocumentation
ArXiv MCP
Overview
ArXiv MCP provides seamless access to arXiv, the premier preprint server for physics, mathematics, computer science, quantitative biology, quantitative finance, statistics, electrical engineering, and systems science. AI agents can search, retrieve, and analyze academic papers directly.
Features
- Paper Search: Search arXiv by title, author, abstract, or category
- Paper Retrieval: Get full paper metadata including abstract, authors, categories, and links
- PDF Access: Direct links to PDF versions of papers
- Category Filtering: Search within specific arXiv categories
- Author Search: Find all papers by a specific author
- Recent Papers: Get the latest submissions in any category
Installation
npx -y @modelcontextprotocol/server-arxiv
Configuration
Add to your Claude Desktop config:
{
"mcpServers": {
"arxiv": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-arxiv"]
}
}
}
No API key required - arXiv API is completely free and open.
Available Tools
| Tool | Description |
|---|---|
search | Search arXiv papers by query |
get_paper | Retrieve a specific paper by ID |
get_recent | Get recent submissions in a category |
get_categories | List all arXiv categories |
Usage Examples
Search for Papers
Search arXiv for: "multi-agent reinforcement learning"
Get Specific Paper
Get paper with ID: "2305.12345"
Recent Papers in Category
Get recent papers in category: "cs.AI" (Artificial Intelligence)
Author Search
Search for papers by author: "Yann LeCun"
Category-specific Search
Search in category "cs.LG" (Machine Learning) for: "transformer attention mechanism"
Claude Desktop Setup
- Install the MCP server
- Add configuration to
claude_desktop_config.json - Restart Claude Desktop
- The ArXiv tool will appear in your available tools
Pros
- ✅ Completely free, no API key required
- ✅ Access to 2+ million preprints
- ✅ Covers all major STEM fields
- ✅ Regular updates with new submissions
- ✅ Well-structured metadata
- ✅ Direct PDF access
Cons
- ❌ Preprints are not peer-reviewed
- ❌ Search API can be slow for complex queries
- ❌ Limited to arXiv content only
- ❌ No full-text search (abstracts only)