import pandas as pd
from fastmcp import FastMCP
from agentic_investor.services.tool_service import ToolService
from agentic_investor.utils.middleware import RequestLoggingMiddleware
from agentic_investor.tools.crypto_fear_greed import CryptoFearGreedTool
from agentic_investor.tools.google_trends import GoogleTrendsTool
from agentic_investor.tools.market_movers import MarketMoversTool
from agentic_investor.tools.cnn_fear_greed import CNNFearGreedTool
from agentic_investor.tools.ticker_data import TickerDataTool
from agentic_investor.tools.options import OptionsTool
from agentic_investor.tools.price_history import PriceHistoryTool
from agentic_investor.tools.financial_statements import FinancialStatementsTool
from agentic_investor.tools.earnings_history import EarningsHistoryTool
from agentic_investor.tools.insider_trades import InsiderTradesTool
from agentic_investor.tools.institutional_holders import InstitutionalHoldersTool
from agentic_investor.tools.nasdaq_earnings_calendar import NasdaqEarningsCalendarTool
from agentic_investor.tools.intraday_data import IntradayDataTool
from agentic_investor.tools.technical_indicators import TechnicalIndicatorsTool
mcp = FastMCP(
"Agentic-Investor",
dependencies=["yfinance", "pandas", "pytrends"],
instructions="""
Use this MCP server for financial and market-related questions, including:
- General questions about stocks, companies, and market performance
- Specific ticker inquiries (e.g., "tell me about Tesla", "how is Apple doing")
- Market analysis and trends
- Questions about equities, bonds, or other market instruments
When a user asks about a company or stock (even generally), use the MCP tools to provide current data.
This server provides real-time and historical financial data including:
- Market movers (gainers, losers, most active stocks)
- Company fundamentals (metrics, news, analyst recommendations)
- Options chains with advanced filtering
- Price history and earnings data
- Financial statements (income, balance sheet, cash flow)
- Institutional ownership and insider trading
- Earnings calendars
- Market sentiment indicators (CNN Fear & Greed, Crypto Fear & Greed, Google Trends)
- Technical indicators (SMA, EMA, RSI, MACD, Bollinger Bands)
- Intraday 15-minute bars via Alpaca API
Recommended workflow when asked about a company/stock:
1. Use get_ticker_data for overview, metrics, recent news, and analyst recommendations
2. Use get_price_history for recent performance trends
3. Add get_financial_statements for fundamental analysis (if relevant)
4. Include get_insider_trades and get_institutional_holders for ownership signals (if relevant)
5. Use calculate_technical_indicator for technical analysis (if relevant)
6. Check market_movers and sentiment indicators for broader context (if relevant)
Always provide current data using these tools rather than relying solely on training knowledge for market-related queries.
All data is cached and optimized for performance. Be concise but thorough in your analysis.
""",
)
# Add request logging middleware
mcp.add_middleware(RequestLoggingMiddleware())
# Initialize tool service and register tools
tool_service = ToolService()
tool_service.register_tools(
[
CryptoFearGreedTool(),
GoogleTrendsTool(),
MarketMoversTool(),
CNNFearGreedTool(),
TickerDataTool(),
OptionsTool(),
PriceHistoryTool(),
FinancialStatementsTool(),
EarningsHistoryTool(),
InsiderTradesTool(),
InstitutionalHoldersTool(),
NasdaqEarningsCalendarTool(),
IntradayDataTool(),
TechnicalIndicatorsTool(),
]
)
# Configure pandas
pd.set_option("future.no_silent_downcasting", True)
# Register tools with MCP
tool_service.register_mcp_handlers(mcp)
if __name__ == "__main__":
mcp.run(transport="http", host="0.0.0.0", port=8000, path="/mcp")
