Arxiv MCP Server

by cyanheads Your server? Claim it

answering

Arxiv MCP Server is answering right now. Last checked 4 min ago. 2 273 installs a week from npm. It exposes 4 tools. Last commit 29 Oct 2026.

Search arXiv, fetch paper metadata, and read full-text content.

Installs per day peak 559 · avg 145 · +197% w/w

a month agotoday

Uptime history 91 days of history · worst day 0%

91 days agonow

100.0%

Uptime 24h

92 of 92 checks

Tools

read from the server

604 ms

Response time

average over 24h

2 273

Installs / week

npm and PyPI

What changed 25

Every tool that appeared, vanished or quietly changed what it asks for. Recorded since 22 August 2026. No other catalogue keeps this.

29 Oct 2 tools changed the parameters they ask for arxiv_get_metadata, arxiv_search_papers

29 Oct a tool changed version

26 Oct a tool changed version

25 Oct a tool changed version

24 Oct a tool changed the parameters it asks for arxiv_search_papers

24 Oct a tool changed version

21 Oct a tool changed version

19 Oct a tool changed version

17 Oct 3 tool descriptions were rewritten arxiv_get_metadata, arxiv_list_categories, arxiv_read_paper

17 Oct 2 tools changed the parameters they ask for arxiv_get_metadata, arxiv_read_paper

and 11 more, back to 22 August 2026

What the code does

We read the source, 3 d ago · tools taken from the live server · rules 995398eba69f

Capabilities

What this server is able to do. For an MCP server this is often the job itself — a terminal server runs commands because that is what it is for. Listed so you know what you are plugging in, not as an accusation.

Runs an external command src/services/arxiv/mirror/refresh-subprocess.ts:110, src/services/arxiv/mirror/query.ts:262

      child = spawn(process.execPath, [entry], {

Is this your server and something here is wrong? Tell us — corrections are free and do not require a plan.

This code can reach further than it looks

We found places where it runs commands, builds paths or queries from values it is given. None of that is a flaw by itself — it becomes one when the code changes, and code changes quietly between releases. We re-read it on every one.

Connect this server

Endpoint below is the one we actually reach during checks — not the one copied from a README. Last verified 4 min ago.

Claude Code Claude Desktop Codex Cursor VS Code

run in your terminal

claude mcp add arxiv-mcp-server --transport http https://arxiv.caseyjhand.com/mcp

Settings → Connectors → Add custom connector, then paste this address

https://arxiv.caseyjhand.com/mcp

~/.codex/config.toml

[mcp_servers.arxiv-mcp-server]
url = "https://arxiv.caseyjhand.com/mcp"

.cursor/mcp.json

{
  "mcpServers": {
    "arxiv-mcp-server": {
      "url": "https://arxiv.caseyjhand.com/mcp"
    }
  }
}

.vscode/mcp.json

{
  "servers": {
    "arxiv-mcp-server": {
      "type": "http",
      "url": "https://arxiv.caseyjhand.com/mcp"
    }
  }
}

Available tools 4

Read directly from the server with tools/list, grouped by what they act on. If a tool disappears, we record the date.

arxiv

arxiv_get_metadata

Get full metadata for one or more arXiv papers by ID. Use when you have known IDs from citations, prior search results, or memory. Looks up 10 IDs per call; any further IDs come back in remaining_ids for the next call.

arxiv_list_categories

List arXiv category codes and names. Useful for discovering valid category filters for arxiv_search_papers. Lists subject classes only; arxiv_search_papers also accepts a bare archive code (the part before the dot, e.g. "astro-ph" or "cs") to search a whole archive at once.

arxiv_read_paper

Fetch the full text of an arXiv paper. Tries arxiv.org/html first, falls back to ar5iv.labs.arxiv.org, and falls back again to text extracted from the PDF when neither has an HTML render — check the source field to know which one answered. Page through long papers with start and max_characters, or pass max_characters null to get the entire body in one call. To read one section, take the outline from the first page (start 0) and pass an entry's start.

arxiv_search_papers

Search arXiv papers by query with category and sort filters. Returns paper metadata including title, authors, abstract, categories, and links.

Tools removed

Tools this server used to expose. Anything built against them stopped working on the day they went.

arxiv_search

removed 17 Oct 2026

Endpoints

URL	Transport	State	Latency	Checked
https://arxiv.caseyjhand.com/mcp	streamable-http	answering	522 ms	4 min ago

Alternatives to Arxiv MCP Server

same job, measured the same way

Project Gutenberg MCP

by bobbyhouse

Read Project Gutenberg books via MCP. Search, fetch metadata, and retrieve text in passages.

local only

SkillDB

by devchad-cmd

Search and load AI agent skills. Free metadata; full content requires a Pro or Studio account.

229 installs/wk answering

Paperport

by cloudcorprecords

Search and read arXiv papers, with local paper storage.

local only

Research papers

by fourzerotwo

arXiv, PubMed, Crossref and OpenAlex scholarly metadata.

38 tools answering

Arxiv MCP Server

by blazickjp

Search arXiv papers, download full text, semantic search, citation graphs, and alerts via MCP.

22 071 installs/wk local only

Reddit MCP Server

by jordanburke

Reddit MCP server with full read/write - posts, comments, search, and content creation.

2 233 installs/wk local only

Raconteur AI Gateway

by getbotlens

MCP gateway for raconteur.com: articles, search and full text. Read-only, no auth.

6 tools answering

Pageport

by cloudcorprecords

Fetch a URL and return its contents as readable markdown text.

230 installs/wk local only

Arxiv MCP Server — questions

Answers built from our own checks of this server.

What can Arxiv MCP Server do?

It exposes 4 tools, read directly from the server on our last check. Among them: arxiv_get_metadata, arxiv_list_categories, arxiv_read_paper, arxiv_search_papers. The full list with descriptions is on this page — we take it from the server itself via tools/list, not from a README. How MCP servers expose tools in the first place →

Is Arxiv MCP Server working right now?

We send a real MCP handshake every 15 minutes. Over the last 24 hours 92 of 92 checks got a reply (100.0%), average response time 604 ms. The bar chart above shows every period we have measured.

Did Arxiv MCP Server ever remove tools?

Yes. arxiv_search is no longer exposed — we recorded the date each one disappeared. A tool vanishing usually means a breaking change for anything that depended on it.

How do I connect Arxiv MCP Server?

Copy the ready config from this page. We generate it for Claude Code, Claude Desktop, Codex, Cursor and VS Code, each in the form that client actually reads. It is a remote server, so there is nothing to install — the client connects to the address.

Does Arxiv MCP Server need an API key?

No. Arxiv MCP Server completed a full MCP handshake with us as an anonymous client and listed its tools without asking for anything. All 4 of them are readable on this page. This is what we observed, not what the docs claim.

How fast is Arxiv MCP Server?

It answers our handshake in 604 ms on average, which is faster than 18% of all working MCP servers we measure. That is on the slow side — worth knowing if the tool sits inside an interactive loop. The comparison comes from our own checks across the whole registry, every 15 minutes.

How many people use Arxiv MCP Server?

The npm package @cyanheads/arxiv-mcp-server was installed 2 273 times in the last week. Week over week that is +197%. We show installs rather than GitHub stars on purpose: a star is a bookmark, an install is someone actually running it.

Is Arxiv MCP Server open source?

The code is open source: it is published under the Apache-2.0 licence, written in TypeScript, 7 stars on GitHub and 4 open issues. The licence covers the code in the repository, not the hosted service at arxiv.caseyjhand.com, which runs on its operator's own terms.