Professional Data MCP Server

A professional-grade Model Context Protocol (MCP) server that provides AI assistants with real-time access to financial, academic, and global development data.

Built with Express.js and TypeScript, supporting both SSE (legacy clients like Claude Desktop) and Streamable HTTP (modern clients like Cursor and Llama.cpp).

---

🚀 Features & Tools

📈 Financial Data (Yahoo Finance)

Tool	Description
finance_get_quote	Real-time stock/ETF/index quotes (Price, Change, PE, etc.)
finance_search	Search for ticker symbols by company name
finance_get_historical	Historical price data (OHLCV) for trend analysis
finance_get_financials	Key financial metrics (Revenue, Debt, ROE, EBIDTA)

📚 arXiv (Advanced Research)

Tool	Description
arxiv_search_papers	Advanced search with field prefixes (ti:, au:, cat:) and date filtering
arxiv_get_paper_data	Full bibliographic metadata for a specific arXiv ID
arxiv_get_full_paper_text	Experimental: Downloads and extracts text content from PDF papers
arxiv_list_categories	Browse research categories (AI, Machine Learning, etc.)
arxiv_update_categories	Scrapes the latest taxonomy directly from arXiv.org

🎓 Academic & Scientific Data

Tool	Description
academic_search_works	Search millions of papers via OpenAlex
academic_get_topic_summary	Get key research insights for a specific topic
scholar_search_papers	Search Semantic Scholar with AI-generated TL;DR summaries
crossref_search_works	Search and filter academic works via Crossref metadata

🌍 Global Development (World Bank)

Tool	Description
worldbank_get_indicator	Fetch global development indicators (GDP, Inflation, etc.)
worldbank_search_indicators	Find specific data points within the World Bank database

🛠 Utilities

Tool	Description
utilities_get_current_time	Get real-time system date/time (Essential for AI time-awareness)

---

🛠 Installation & Setup

Local Development

1. Clone the repository
2. Install dependencies:

   npm install

3. Build and Start:

   npm run build
   npm start

The server will listen on http://localhost:3000.

---

🔌 Connection Guide

1. Cursor / Llama.cpp (Streamable HTTP)

Use the following endpoint URL. Note: Cursor requires a dummy query parameter.

http://localhost:3000/mcp?apiKey=123

2. Claude Desktop (SSE)

Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "professional-data": {
      "url": "http://localhost:3000/sse"
    }
  }
}

---

☁️ Deployment

Railway.app

This project is ready for one-click deployment on Railway.

1. Connect your GitHub repository to Railway.
2. The Procfile and Dockerfile are already configured.
3. Ensure the PORT environment variable is set to 3000.

---

📝 License

Distributed under the MIT License. See LICENSE for more information.