yzfly/arxiv-mcp-server

This page was translated automatically. Technical details may differ from the original. View the original

44 stars 4 forks Quality 76 Trend 16MIT

🔬 ArXiv Paper Intelligent Interpretation Assistant - Arxiv-MCP-Server supports one-click download, parsing, and translation of academic papers under the MCP protocol into Chinese, and generates WeChat official account article formats

Overview

An arXiv-based paper interpretation and knowledge organization assistant. Supports Model Context Protocol (MCP) standards, converting academic papers into plain Chinese interpretation and WeChat official account articles with one click, suitable for personal use or integration with automated tools. * 🔍 : Keyword search for quickly locating papers you care about * 📥 : Automatically retrieve and save original papers * 📝 : Parse PDF English content into high-quality Chinese Markdown for easy note-taking, archiving, and reuse * 📱 : Automatically generate popular article drafts adapted for WeChat reading experience * 🗑️ : Supports one-click clearing of all historical files to avoid storage backlog * 🤖 : Powered by SiliconFlow large model services, content processing is efficient and natural Please obtain the SiliconFlow API Key via the following link: https://cloud.siliconflow.cn/i/TxUlXG3u Configuration file location: - : ~/Library/Application Support/Claude/claude_desktop_config.json - : %APPDATA%\Claude\claude_desktop_config.json * input: arXiv paper URL or arXiv ID (e.g., 2403.15137v1) * arxivId: arXiv paper ID * paperInfo: Meta information of the paper (optional, including title/author/abstract, etc.) * arxivId: arXiv thesis ID * includeWechat: Whether to generate WeChat articles (optional, default true is true) * {arxivId}.pdf - Original PDF * {arxivId}_text.txt - English original parsed text * {arxivId}_md_zh.md - Chinese Markdown interpretation * {arxivId}_wechat.md - WeChat official account article * (You can use tools to directly get these filenames)

Overview (machine translated)

Feature Highlights

🔍 arXiv Paper Intelligent Search: Keyword search to quickly locate the papers you care about
📥 One-click PDF download: Automatically retrieves and saves original papers 📝 Intelligent Chinese-English Conversion: Parses PDF English content into high-quality Chinese Markdown for easy note-taking, archiving, and reuse
📱 WeChat Article Generation: Automatically generates popular article drafts tailored to the WeChat reading experience
🗑️ One-Click File Cleanup: Supports one-click clearing of all historical files to avoid space backlog
🤖 AI Content Understanding and Rewriting: Leveraging SiliconFlow large model services, content processing is efficient and natural

Each file processing tool returns the actual saved file name, making it easy to integrate into any automated workflow!

Installation and Use

NPX Methods (Recommended)

Global Installation

Configuration Requirements

Environment variables

Before use, please set the following environment variables:

Obtaining the API Key

Please obtain the SiliconFlow API Key via the following link: https://cloud.siliconflow.cn/i/TxUlXG3u

MCP Client Configuration

Claude Desktop configuration

Add in the Claude Desktop configuration file:

Configuration file location:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

Other MCP Clients

For other clients that support MCP, please refer to their documentation for configuring stdio transfer methods.

Available Tools and Parameters

After each tool is processed, it returns the saved file name, making it easy to track or download later.

1. Search for papers

ToolName: search_arxiv
Parameters:
query: Keyword
maxResults: Number of papers returned (optional, default 5)

2. Download PDF

Tool Name: download_arxiv_pdf
Parameters:
input: arXiv paper URL or arXiv ID (e.g., 2403.15137v1)

3. Parsed as Chinese Markdown

Tool Name: parse_pdf_to_markdown
Parameters:
arxivId: arXiv paper ID

This page was translated automatically. Technical details may differ from the original.

Install

npx -y @langgpt/arxiv-mcp-server@latest

New to MCP? Start with the mcpskill.org guide →

Configuration

{
  "mcpServers": {
    "arxiv-mcp-server": {
      "command": "npx",
      "args": ["-y", "@langgpt/arxiv-mcp-server@latest"],
      "env": {
        "SILICONFLOW_API_KEY": "your_api_key_here",
        "WORK_DIR": "/path/to/your/data/directory"
      }
    }
  }
}