This project provides a Model Context Protocol (MCP) server that bridges AI assistants with the arXiv research repository, enabling programmatic searching and access to academic papers. It is designed for AI developers and researchers looking to integrate arXiv data into their AI workflows, offering efficient paper discovery and content retrieval.

How It Works

The server leverages the Model Context Protocol (MCP) to expose functionalities for searching, downloading, listing, and reading arXiv papers. Papers are stored locally for faster subsequent access. It supports filtering searches by date ranges and categories, and provides a structured workflow for deep paper analysis via specialized prompts.

Quick Start & Requirements

• Install via Smithery: npx -y @smithery/cli install arxiv-mcp-server --client claude
• Manual install: uv tool install arxiv-mcp-server
• Development install: Requires cloning the repository, setting up a virtual environment with uv, and installing with uv pip install -e ".[test]".
• Prerequisites: Python, uv (or pip), npx (for Smithery).
• Storage path can be configured via ARXIV_STORAGE_PATH environment variable.
• Official documentation and contribution guidelines are linked in the README.

Highlighted Details

• Enables AI assistants to interact with arXiv papers via MCP.
• Supports paper search with filters (date, categories) and content access.
• Includes a deep-paper-analysis prompt for comprehensive paper evaluation.
• Local storage for downloaded papers enhances access speed.

Maintenance & Community

• Developed by the Pearl Labs Team.
• Contribution guidelines and bug reporting are available via GitHub links.

Licensing & Compatibility

• Released under the MIT License.
• Permissive license suitable for commercial use and integration into closed-source projects.

Limitations & Caveats

The project appears to be in active development, with installation primarily targeting users familiar with uv and MCP clients. Specific performance benchmarks or detailed system requirements beyond standard Python environments are not provided.