arxiv-mcp
Enables searching arXiv and downloading papers via MCP tools, no API key required.
README
arxiv-mcp
An MCP server that lets any coding agent search arXiv and download papers. Free, no API key, two tools.
Tools
| Tool | What it does |
|---|---|
search_arxiv(query, max_results=10) |
Free-text search. Returns title, authors, abstract, date, PDF URL. No PDF parsing. |
download_paper(arxiv_id) |
Streams a PDF into ./papers/<id>.pdf. Idempotent. |
Both are read-only on the network side. The server touches arXiv's public
API (export.arxiv.org) — no key, no signup, no rate-limit token.
Install
git clone https://github.com/vishal24p/arxiv_mcp.git
cd arxiv_mcp
uv venv
uv pip install -e .
Requires Python 3.10+ and uv.
Wire it into your agent
The arxiv-mcp script is installed into the venv at .venv\Scripts\
(Windows) or .venv/bin/ (macOS/Linux). Activate first, or call it by
path:
# Activate (then `arxiv-mcp` is on PATH)
# Windows (PowerShell):
.venv\Scripts\Activate.ps1
# Windows (cmd):
.venv\Scripts\activate.bat
# macOS / Linux:
source .venv/bin/activate
arxiv-mcp install # auto-detect and patch all found agents
arxiv-mcp install --only codex # one specific agent
arxiv-mcp install --dry-run # show what would change, write nothing
If you skip activation, invoke the executable directly:
# Windows
.venv\Scripts\arxiv-mcp.exe install
# macOS / Linux
.venv/bin/arxiv-mcp install
Supported agents:
| Agent | Config file | Format |
|---|---|---|
| Claude Desktop | %APPDATA%\Claude\claude_desktop_config.json |
JSON |
| Claude Code | ~/.claude.json |
JSON |
| Cursor | ~/.cursor/mcp.json |
JSON |
| Codex CLI | ~/.codex/config.toml |
TOML |
The installer merges an arxiv entry into each detected config, backs up
the original to <file>.bak, and never touches other settings. It is
idempotent — re-running with the same flag reports the entry already
exists and writes nothing. Restart the agent and the tools appear.
Use
Ask the agent anything that needs arXiv:
find 3 papers on mixture of experts
download 1706.03762
what's the most cited recent paper on RLHF?
The agent picks search_arxiv or download_paper based on what you ask.
PDFs land in ./papers/. Override the location with
ARXIV_MCP_PAPERS_DIR.
Run without installing to an agent
arxiv-mcp serve
# ...or:
python -m arxiv_mcp
Speaks MCP over stdio. Useful for testing or wiring into something that isn't a coding agent.
Troubleshooting
arxiv-mcp is not recognized as the name of a cmdlet / command
The script lives in the venv at .venv\Scripts\arxiv-mcp.exe (Windows)
or .venv/bin/arxiv-mcp (macOS/Linux). It is only on PATH while the
venv is active. Either activate the venv first, or invoke the
executable by its full path (see Wire it into your agent above).
The script name uses a hyphen (arxiv-mcp), not an underscore
(arxiv_mcp). The Python package is arxiv_mcp; the console script is
arxiv-mcp.
Re-running arxiv-mcp install reports the entry already exists
That is the intended idempotent behavior. Backups are at
<config>.bak next to each original file. To force a fresh install,
delete the arxiv key from the relevant config and re-run.
License
Apache 2.0.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.