MCP Catalogs
Home

arxiv-mcp-server

by blazickjp·2,738·Score 58

An MCP server for searching and analyzing ArXiv papers with local storage and advanced search capabilities.

ai-llmsearchdeveloper-tools
220
Forks
16
Open issues
1 mo ago
Last commit
2d ago
Indexed

Overview

The ArXiv MCP Server provides a bridge between AI assistants and ArXiv's research repository through the Model Context Protocol. It allows AI models to search for papers by various criteria including date ranges and categories, download paper content for offline access, and maintain a local library of research papers. The server includes specialized prompts for academic paper analysis and implements security measures to address prompt injection risks from untrusted external content.

Try asking AI

After installing, here are 7 things you can ask your AI assistant:

you:AI research assistants searching and summarizing academic papers
you:Researchers staying updated on latest publications in their field
you:Automated literature reviews for meta-analyses
you:Content creators curating research summaries for audiences
you:Does this server handle PDF papers?
you:How does the server handle ArXiv rate limits?
you:Can I use this server with MCP clients other than Claude?

When to choose this

Choose this server for research-focused AI applications that need access to academic papers from arXiv, with proper search capabilities and local caching.

When NOT to choose this

Don't choose this if you need write access to papers, as it only provides read-only access, or if you require processing papers from other academic repositories.

Tools this server exposes

9 tools extracted from the README
  • search_papers

    Search arXiv with optional category, date, and boolean filters

  • download_paper

    Download a paper by its arXiv ID for offline access

  • list_papers

    List all papers downloaded locally

  • read_paper

    Read the full text of a locally downloaded paper in markdown

  • semantic_search

    Semantic similarity search over locally downloaded papers

  • deep-paper-analysis

    Comprehensive workflow for analyzing academic papers

  • summarize_paper

    Generate a concise structured summary of a paper

  • compare_papers

    Side-by-side technical comparison across multiple paper IDs

  • literature_review

    Thematic synthesis across a topic and optional paper set

Comparable tools

semantic-search-mcpbrowser-mcpfile-system-mcp

Installation

Installing via Smithery

npx -y @smithery/cli install arxiv-mcp-server --client claude

Installing via Claude Desktop (.mcpb)

  1. Download the matching artifact from the [latest release](https://github.com/blazickjp/arxiv-mcp-server/releases/latest)
  2. In Claude Desktop open **Settings → Extensions** (or drag-and-drop)
  3. Click **Install** and set your paper storage directory

Manual Installation

uv tool install arxiv-mcp-server

Claude Desktop Configuration

{
    "mcpServers": {
        "arxiv-mcp-server": {
            "command": "uv",
            "args": [
                "tool",
                "run",
                "arxiv-mcp-server",
                "--storage-path", "/path/to/paper/storage"
            ]
        }
    }
}

FAQ

Does this server handle PDF papers?
Yes, by default it handles HTML versions. For PDF-only older papers, install with the [pdf] extra: 'uv tool install "arxiv-mcp-server[pdf]"'
How does the server handle ArXiv rate limits?
The server automatically enforces ArXiv's 3-second rate limit. If rate limited, it advises waiting 60 seconds before retrying.
Can I use this server with MCP clients other than Claude?
Yes, the server supports stdio transport and Streamable HTTP for broader MCP client compatibility.

Compare arxiv-mcp-server with

GitHub →

Last updated · Auto-generated from public README + GitHub signals.