IndiaQuant MCP – AI Stock Market Assistant

IndiaQuant MCP is a real-time Indian stock market assistant built using Python and Model Context Protocol (MCP). It integrates with Claude Desktop to provide live stock market intelligence such as stock prices, trading signals, options data, sentiment analysis, and portfolio tracking.

The system uses only free APIs and open-source libraries.

Project Overview

Normally AI assistants like Claude do not have access to live stock market data.
This project solves that problem by building an MCP server that connects Claude to a stock market backend.

Architecture:

Claude Desktop

MCP Server (mcp_server.py)

FastAPI Backend (main.py)

Free Market APIs (yfinance, NewsAPI, Alpha Vantage)

Claude can then call the MCP tools and receive real-time market data.

Features

This project implements the following capabilities:

1. Fetch live stock prices from NSE

2. Generate AI trading signals (BUY / SELL / HOLD)

3. Analyze options chain data

4. Calculate option Greeks (Delta, Gamma, Theta, Vega)

5. Detect unusual options activity

6. Scan the market for opportunities

7. Analyze news sentiment

8. Show sector performance heatmap

9. Maintain a virtual trading portfolio

10. Calculate real-time portfolio profit and loss

All tools return live market data using free APIs.

Technologies Used

Python
FastAPI
yfinance
pandas
numpy
pandas-ta
SQLite
Model Context Protocol (MCP)
Claude Desktop

Project Structure

indiaquant-project

main.py # FastAPI backend server mcp_server.py # MCP server exposing tools to Claude requirements.txt # Python dependencies README.md # Project documentation

Installation

cd indiaquant-mcp Install dependencies

pip install -r

requirements.txt

fastapi uvicorn yfinance pandas numpy pandas-ta requests mcp scipy sqlite-utils python-dotenv newsapi-python alpha_vantage

Running the Backend

Start the FastAPI server: uvicorn main:app --reload

This starts the API that provides market data and analysis.

MCP Configuration

To connect the project with Claude Desktop, create the following configuration file:

C:\Users\MY_USERNAME\AppData\Roaming\Claude\claude_desktop_config.json

Add this configuration:

{
  "mcpServers": {
    "indiaquant": {
      "command": "C:/Users/USERNAME/AppData/Local/Programs/Python/Python314/python.exe",
      "args": [
        "C:/Users/USERNAME/Desktop/indiaquant-project/mcp_server.py"
      ]
    }
  }
}

After adding the configuration, restart Claude Desktop.

Claude will automatically start the MCP server.

Example Queries in Claude

After connecting the MCP server, Claude can call the tools.

Example queries:

Use indiaquant to get live price of RELIANCE.NS
Use indiaquant to generate signal for HDFCBANK.NS
Use indiaquant to detect unusual options activity on INFY.NS
Use indiaquant to scan the market

Claude will call the MCP tools and return live market data.

APIs Used

This project only uses free APIs.

yfinance – live stock prices and options chain
NewsAPI – news sentiment analysis
Alpha Vantage – macro indicators
pandas-ta – technical indicators

No paid APIs are used.

Design Decisions

FastAPI was chosen for building a lightweight backend API.

yfinance provides free access to NSE stock data.

SQLite is used for storing virtual portfolio data because it is simple and portable.

The MCP layer allows Claude to communicate with the backend using tools.

Future Improvements

Possible future enhancements include:

• Deploy the server to cloud for 24/7 availability
• Add Redis caching for faster market data access
• Improve trading signals using machine learning
• Integrate real brokerage APIs for real trading