arXiv MCP Server
Search, download and read arXiv papers section by section, with BibTeX export.
Community, verifiedby Joe BlazickRuns on your computerChecked October 18, 2026
This community server lets Claude search arXiv, download papers to your computer and read them one section at a time, including the authors' original LaTeX. It also exports BibTeX, follows references and citing papers through Semantic Scholar, and keeps topic watches that flag new papers.
Who it is for
- AI and science researchers
- Graduate students writing literature reviews
- Engineers keeping up with new papers
Prompts to try
Once it is connected, ask in plain words. Claude decides when to use the server and asks before it changes anything important.
Search arXiv for papers on speculative decoding from the last six months and summarize the top three.Download arXiv paper 1706.03762 and explain its method section in plain language.Export BibTeX entries for these four arXiv IDs.How to connect it
In Claude Code
claude mcp add --transport stdio --scope user arxiv -- uvx arxiv-mcp-server
In the Claude desktop app
Add this to claude_desktop_config.json, then restart the app:
"arxiv": {"command": "uvx", "args": ["arxiv-mcp-server"]} (macOS .mcpb bundles are on the GitHub releases page)Good to know
Install from PyPI with uvx, not npm (an unrelated npm package uses the same name); papers are stored in ~/.arxiv-mcp-server/papers.
Unlock the setup steps
Free. Leave your email once to open the setup steps for all 150 servers, and get the best new connectors each month.
How we checked it
Checked the GitHub repo (Apache-2.0, 3.2k stars, release v0.8.2 on 2026-10-08), the arxiv-mcp-server PyPI package and the registry entry io.github.blazickjp/arxiv-mcp-server.