Deep Thinking Trading System 🧠 📈

A sophisticated multi-agent AI system for financial analysis, powered by NVIDIA AI and orchestrated with LangGraph. It features a modern, premium React frontend and a robust FastAPI backend.

Note

This project is inspired by the research paper: TradingAgents: Multi-Agents LLM Financial Trading Framework.

🚀 Features

• Multi-Agent Architecture: Separate agents for Market Data, News, Social Sentiment, Fundamentals, Research, Trading, and Risk Management.
• Adversarial Debates: Bull vs. Bear and Risk Management debates to stress-test investment ideas.
• Live Data: Real-time integration with Yahoo Finance, Finnhub, Tavily Search, Alpha Vantage, and Financial Datasets.
• NVIDIA Powered: Uses NVIDIA-hosted LLMs (via OpenAI-compatible API) for deep reasoning and fast data processing.
• Premium UI: Dark-themed, glassmorphic React dashboard to visualize the agent's thought process.

🛠️ Tech Stack

• Backend: Python, FastAPI, LangGraph, LangChain, NVIDIA AI API
• Frontend: React, Vite, Vanilla CSS (Premium Design), Lucide Icons
• Data Sources: Yahoo Finance, Finnhub, Tavily, Alpha Vantage, Financial Datasets

📦 Installation

Prerequisites

• Python 3.10+
• Node.js 16+
• API Keys (copy .env.example to .env and fill in your keys):
  • NVIDIA_API_KEY — NVIDIA Build (required)
  • TAVILY_API_KEY — Tavily (required for web search)
  • FINNHUB_API_KEY — Finnhub (required for financial news)
  • ALPHA_VANTAGE_API — Alpha Vantage (required for Indian/global stock data)
  • FINANCIAL_DATASETS_API_KEY — Financial Datasets (optional)

1. Backend Setup

1. Navigate to the root directory.
2. Create virtual environment (optional but recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

   source server/venv/bin/activate
   pip install -r server/requirements.txt
   python server/main.py

   Or directly:

   server/venv/bin/python server/main.py

   The API will be available at http://localhost:8000.

2. Frontend Setup

1. Open a new terminal and navigate to the client folder:

   cd client

2. Install dependencies:

   npm install

3. Run the Development Server:

   npm run dev

   Access the UI at http://localhost:5173.

🏗️ Architecture

The system follows a Client-Server architecture:

1. Client: The React app sends a request with a Ticker Symbol to the FastAPI backend.
2. Server: Triggered by the API, langgraph initiates the workflow:
   • Analyst Team: Gathers raw data.
   • Research Team: Bull and Bear agents debate the data; Manager synthesizes a plan.
   • Trader: Proposes a trade execution.
   • Risk Team: Debates the safety of the trade; Portfolio Manager makes the final decision.
3. Response: The final decision and all intermediate reports are sent back to the Client for display.

📝 License

MIT