arxiv-search

Provides arXiv search and paper retrieval via an MCP server for easy agent integration.

python

2

GitHub Stars

python

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "markmcd-arxiv-search-mcp": {
      "command": "python",
      "args": [
        "-m",
        "arxiv_mcp"
      ]
    }
  }
}

You can run an MCP server that lets your agent search arXiv and fetch paper details by starting a local Python-based server and exposing its capabilities to your MCP client. This server provides two core actions: searching papers on arXiv and retrieving detailed information for a specific paper.

How to use

To use the arXiv MCP server with your MCP client, you can perform two main actions. First, search for papers with a query to discover relevant results. Second, request detailed information about a specific paper by providing its identifier. You simply invoke the server’s endpoints through the MCP client as you would with other MCP servers, using the tool names described below. The server handles the arXiv queries and returns structured results that you can present or filter in your workflow.

How to install

Prerequisites: you need Python and pip available on your system.

Install the MCP server package from PyPI.

Install the package in your environment using the following command.

pip install arxiv-search-mcp

Additional setup and running

Load and configure the server for your MCP client. You can start the server locally with Python and expose it to your MCP workflow.

To run the server module directly from the command line.

python -m arxiv_search_mcp

Configure the server in your MCP client

Include the arXiv MCP server in your agent configuration by specifying the command to launch the server and its arguments.

{
  "mcpServers": {
    "arxiv": {
      "command": "python",
      "args": ["-m", "arxiv_mcp"]
    }
  }
}

Available tools

search_papers

Search for papers on arXiv using a query string and optional maximum number of results.

get_paper

Retrieve detailed information for a specific arXiv paper by its ID.