---
id: arxiv-mcp-server
version: "0.6.3"
license: Apache-2.0
license_treatment: permissive
maintenance: active
---

# arxiv-mcp-server — A flexible arXiv search and analysis service with MCP protocol support

License: permissive · Maintenance: active · Downloads: 83.2K/mo

## What it is and what it does

arxiv-mcp-server is an MCP (Model Context Protocol) server that bridges AI assistants like Claude and Codex to arXiv research infrastructure. It exposes 14 tools for searching papers by query and category, downloading and reading full text locally, retrieving author-submitted LaTeX by section, traversing citation graphs via Semantic Scholar, exporting BibTeX metadata, and maintaining persistent research alerts. Papers and indexes stay on your machine; searches and external API calls go to arXiv and Semantic Scholar.

The server runs locally over stdio by default and integrates into Claude Desktop, VS Code, Kiro, and other MCP clients via configuration. It supports command-line installation through uv, plugin marketplaces for Claude Code and Codex, and a bundled macOS extension. The package depends on aiohttp, httpx, pydantic, starlette, and the arxiv Python client to coordinate these operations.

Use it for:

- Search arXiv by keyword or category, fetch abstracts, and download papers into a local library for an AI assistant to analyze.
- Retrieve bounded LaTeX sections from papers to examine mathematical proofs or algorithm details without downloading full PDFs.
- Follow citation graphs to discover related work: fetch references cited by a paper or papers that cite it.
- Set up research alerts on specific arXiv topics and check for new papers matching your watch list since the last check.
- Export BibTeX for multiple papers at once to build a bibliography for a research project.
- Search downloaded papers by semantic similarity (with [pro] extra) to find related work in your local library.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

An MCP server that exposes arXiv search, paper download, full-text reading, LaTeX retrieval, citation graph traversal, and research alert management as tools for AI clients like Claude and Codex.

Yes, if you use Claude, Codex, or another MCP-compatible AI client and want to integrate arXiv research into your workflow. The package is actively maintained, has no known vulnerabilities, low install friction, and permissive licensing. Install via uv for command-based clients or use the bundled .mcpb extension for Claude Desktop on macOS. Not suitable if you need a standalone Python library rather than an MCP server.

## Install

pip install arxiv-mcp-server
uv add arxiv-mcp-server
poetry add arxiv-mcp-server

## Installing arxiv-mcp-server

Before you install: Low friction: pure Python wheel with no compiled dependencies. Active maintenance—released 2026-08-14 with 3053 GitHub stars. Requires Python 3.11+ and uv for command-based client integrations; desktop clients can use the bundled .mcpb extension.

License in practice: Apache-2.0 permissive license allows commercial and private use without restriction.

Quickstart:

# Install via uv (recommended for MCP clients)
uv tool install arxiv-mcp-server

# Or run directly with uvx
uvx arxiv-mcp-server

# In Claude Desktop or compatible MCP client, configure:
# {
#   "mcpServers": {
#     "arxiv": {
#       "type": "stdio",
#       "command": "uvx",
#       "args": ["arxiv-mcp-server"]
#     }
#   }
# }

Requires Python 3.11 or later. Desktop clients must have uv in PATH or use absolute path to uvx. PDF conversion for older papers requires installing with [pdf] extra.

Verify before relying:

- Whether semantic search and reindex tools ([pro] features) require a separate subscription or paid tier.
- Performance characteristics when handling large citation graphs or bulk paper downloads.
- Exact storage footprint for a typical research library of downloaded papers.

## Package facts

- License: Apache-2.0 (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 83.2K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

arxiv paper search mcp server, ai research paper retrieval, mcp protocol arxiv integration, academic paper download tool, citation graph exploration, research alert monitoring, arxiv latex source access, mcp-server, research-tools, arxiv-integration

[View on SkillFed](https://skillfed.io/packages/arxiv-mcp-server) · [View on PyPI](https://pypi.org/project/arxiv-mcp-server/)