Repolode
Submit
blazickjp

blazickjp/arxiv-mcp-server

A local MCP server that lets AI agents search arXiv, read papers section by section from original LaTeX, export BibTeX, and watch topics.

View on GitHubWebsite
Stars
3,258
Forks
261
Open issues
7
Last commit
30 days ago

What is it?

arxiv-mcp-server is a local Model Context Protocol server for AI agents doing literature research on arXiv. Its core workflow runs from a paper ID to its outline, then to a single section read from the author-submitted LaTeX source, and finally to BibTeX citations exported from arXiv metadata. Downloaded papers and topic watches are stored on the local disk.

Search is available but optional. The server exposes tools for arXiv queries with category and date filters, abstract retrieval, paper download and conversion to Markdown, bounded paragraph-level reading, passage search, and citation graph lookups through Semantic Scholar. Semantic search over downloaded papers is available as an extra install.

It runs over stdio by default and can be launched with uvx, and it also supports Streamable HTTP with a bound local address. The server includes built-in prompt workflows for paper summaries, comparisons, and literature reviews. Paper text is treated as untrusted content, and the project documents guidance on prompt injection risks.

Key features

  • 01Reads original arXiv LaTeX one section at a time through an outline.
  • 02Exports BibTeX from arXiv metadata.
  • 03Saves topic watches locally and checks them for new papers.
  • 04Converts papers to Markdown with bounded chunked reading.
  • 05Supports arXiv search with category, date, and sort filters.
  • 06Offers optional local semantic search over downloaded papers.

Use cases

  • Drafting a related-work section by reading specific sections of source papers.
  • Generating BibTeX entries for a bibliography from arXiv IDs.
  • Monitoring arXiv categories for new papers on a research topic.
  • Tracing references and citing papers of a seed paper through a citation graph.

Quick start

  1. 1

    Install uv, which provides uvx

  2. 2

    Run the server with uvx

    uvx arxiv-mcp-server
  3. 3

    Add the stdio configuration to your MCP client's mcpServers settings

Things to know

  • Semantic search and reindexing require the optional [pro] extra.
  • The Claude Desktop bundle is for macOS only.
  • An unrelated npm package uses the same name, so do not install this server with npm.

In collections

MCP servers for Claude CodeClaude Code tools, skills & pluginsMCP servers

Topics

#mcp-server#arxiv#research#literature-review#bibtex#llm#claude#latex#ai#claude-ai

FAQ

What is arxiv-mcp-server?
A local MCP server that lets AI agents search arXiv, read papers section by section from original LaTeX, export BibTeX, and watch topics. arxiv-mcp-server is a local Model Context Protocol server for AI agents doing literature research on arXiv. Its core workflow runs from a paper ID to its outline, then to a single section read from the author-submitted LaTeX source, and finally to BibTeX citations exported from arXiv metadata. Downloaded papers and topic watches are stored on the local disk.
Is arxiv-mcp-server free to use? What license does it have?
arxiv-mcp-server is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is arxiv-mcp-server still maintained?
Actively maintained. The last commit to arxiv-mcp-server was on Oct 8, 2026.
What can I use arxiv-mcp-server for?
Drafting a related-work section by reading specific sections of source papers. Generating BibTeX entries for a bibliography from arXiv IDs. Monitoring arXiv categories for new papers on a research topic. Tracing references and citing papers of a seed paper through a citation graph.
How do I install arxiv-mcp-server?
Install uv, which provides uvx Run the server with uvx: uvx arxiv-mcp-server Add the stdio configuration to your MCP client's mcpServers settings
Where can I get arxiv-mcp-server?
The source code and installation instructions are on GitHub: https://github.com/blazickjp/arxiv-mcp-server

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects