Financial Stock Market MCP Server

A Financial MCP Project

A Model Context Protocol (MCP) implementation demonstrating financial data analysis as a use case, built with yfinance and featuring a comprehensive server with tools, resources, and prompts for stock market analysis. Now enhanced with enterprise-grade guardrails for security, compliance, and safety. This project was inspired after I enrolled in the MCP Course offered by DeepLearning.ai in collaboration with Anthropic.

🚀 Features

MCP Server (financial_server.py)

• Tools: 4 powerful financial analysis tools
  • get_stock_info: Get comprehensive stock information
  • get_historical_data: Retrieve historical price data
  • compare_stocks: Compare multiple stocks by metrics
  • get_market_summary: Get major market indices summary
• Resources: Access saved financial data
  • finance://portfolios: List all saved analyses
  • finance://{filename}: Access specific data files
• Prompts: Pre-built analysis templates
  • analyze_stock_prompt: Comprehensive stock analysis
  • portfolio_comparison_prompt: Multi-stock portfolio analysis

Enhanced MCP Client (enhanced_financial_chatbot.py)

• Interactive chat interface with financial focus
• 🛡️ Comprehensive guardrails system
• Support for resources (@portfolios, @filename)
• Support for prompts (/prompts, /prompt )
• Integration with multiple MCP servers
• Real-time financial data queries
• Session tracking and monitoring
• Investment advice detection and blocking

🛡️ Comprehensive Guardrails System (guardrails.py)

• Content Filtering: Blocks investment advice requests and high-risk content
• Rate Limiting: Prevents API abuse (15 calls/minute, 200/hour, 2000/day)
• Input Validation: Stock symbol format validation and sanitization
• Security Protection: Code injection detection and blocking
• Session Management: User session tracking with violation logging
• Risk Assessment: Automatic query risk level evaluation
• Response Enhancement: Automatic disclaimers and safety warnings
• Compliance Features: Regulatory compliance and professional referrals

📦 Installation

1. Clone and setup the project:

# Create project directory
mkdir financial-mcp-project
cd financial-mcp-project

# Initialize with uv
uv init

2. Install dependencies:

# Create virtual environment
uv venv

# Activate virtual environment
source .venv/bin/activate  # Linux/Mac
# or
.venv\Scripts\activate     # Windows

# Install core dependencies
uv add mcp yfinance anthropic python-dotenv nest-asyncio pandas numpy

# Install security libraries for guardrails
uv add ratelimit validators bleach

3. Set up environment variables:

# Copy example environment file
cp .env.example .env

# Edit .env with your Anthropic API key
ANTHROPIC_API_KEY=your_actual_api_key_here

4. Create required files and directories:

# Create data directory
mkdir -p financial_data

# Create enhanced version directory
mkdir -p enhanced_version

# Create guardrails configuration in enhanced_version folder
touch enhanced_version/guardrails_config.json

🏃‍♂️ Quick Start

Running the MCP Server (Testing)

# Test the server with MCP inspector
npx @modelcontextprotocol/inspector uv run financial_server.py

Running the Enhanced Financial Chatbot

# Navigate to enhanced version folder
cd enhanced_version

# Start the chatbot with comprehensive guardrails
uv run enhanced_financial_chatbot.py

Running the Simple Version (Optional)

# Navigate to enhanced version folder
cd enhanced_version

# Start the chatbot with basic guardrails
uv run simple_financial_chatbot.py

Running the Original Basic Chatbot

# From main directory
uv run financial_chatbot.py

💬 Usage Examples

Basic Stock Queries

Query: What's the current price of Apple stock?
Query: Get me historical data for TSLA over the past year
Query: Compare AAPL, GOOGL, and MSFT by market cap

Using Resources

Query: @portfolios                    # List all saved data
Query: @AAPL_info.json               # View specific stock data
Query: @market_summary_20241220.json # View market summary

Using Prompts

Query: /prompts                                              # List available prompts
Query: /prompt analyze_stock_prompt symbol=AAPL            # Analyze Apple stock
Query: /prompt portfolio_comparison_prompt symbols=["AAPL","GOOGL","MSFT"] timeframe=1y

Guardrails Features

Query: /status                           # Check session statistics and violations
Query: Should I buy Apple stock?         # Will be blocked - investment advice
Query: What's a good investment?         # Will be blocked - investment advice

Advanced Analysis

Query: Analyze the tech sector by comparing AAPL