mcp-paperswithcode

Name: mcp-paperswithcode
Author: hbg

@hbg

Repository

About mcp-paperswithcode

An unofficial MCP interface to interact with the PapersWithCode API

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "mcp-paperswithcode": {
      "command": "npx",
      "args": [
        "-y",
        "@smithery/cli",
        "install",
        "@hbg/mcp-paperswithcode",
        "--client",
        "claude"
      ]
    }
  }
}

Tools

No tools detected

Fetch the live tool list by running this server in a temporary sandbox using the button above.

Overview

What is mCP-paperswithcode?

mcp-paperswithcode is a Model Context Protocol (MCP) server that interfaces with the PapersWithCode API to allow AI assistants to find, read, and retrieve information about research papers, authors, datasets, conferences, and linked code repositories.

How to use mcp-paperswithcode?

Install the server automatically for Claude Desktop via Smithery using the command npx -y @smithery/cli install @hbg/mcp-paperswithcode --client claude. Once installed, AI assistants can invoke the provided tools to search and retrieve paper metadata, author details, conference proceedings, research area information, and more.

Key features of mcp-paperswithcode

Search papers by title, abstract, or ArXiv ID.
Retrieve full metadata for a specific paper.
Extract readable text from a paper's PDF or HTML URL.
List benchmarks, tasks, methods, repositories, and datasets for a paper.
Search and browse research areas, authors, and conferences.

Use cases of mcp-paperswithcode

An AI assistant helps a researcher find recent papers on a specific topic and view their benchmark results and linked code repositories.
A developer searches for papers by a known author and retrieves all their published work along with associated methods and datasets.
A conference attendee explores proceedings and lists all papers accepted at a specific conference or proceeding.
A student looks up a paper's details by its ArXiv ID and then reads the full text by providing its URL.