Arxiv MCP Server

@pipeworx/arxiv

arXiv preprint server MCP — full-text search and paper lookup, no auth.

Part of Pipeworx — an MCP gateway connecting AI agents to 1394+ live data sources.

Tools

• search_papers(query, max_results?, sort_by?, sort_order?) — text search; supports field prefixes au:, ti:, abs:, cat:, all:. Combine with AND, OR, ANDNOT.
• get_paper(arxiv_id) — fetch one paper by ID (e.g., 2310.06825).

Data source

https://export.arxiv.org/api/query — public Atom-feed endpoint, no key required.

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "arxiv": {
      "url": "https://gateway.pipeworx.io/arxiv/mcp"
    }
  }
}

Or connect to the full Pipeworx gateway for access to all 1394+ data sources:

{
  "mcpServers": {
    "pipeworx": {
      "url": "https://gateway.pipeworx.io/mcp"
    }
  }
}

Using with ask_pipeworx

Instead of calling tools directly, you can ask questions in plain English:

ask_pipeworx({ question: "your question about Arxiv data" })

The gateway picks the right tool and fills the arguments automatically.

More

• Docs and guides
• pipeworx.io

License

MIT