OptiNewsTrader - AI-Powered News Analysis System

OptiNewsTrader is an advanced, AI-powered web application designed to identify short-term options trading opportunities (0-7 days) based on real-time news analysis, technical indicators, and market data.

Key Features

• Real-time News Ingestion: Minute-level polling from multiple sources (Alpaca, FMP, Searxng)
• AI-Powered Analysis: Deepseek AI with phased workflow (Research → Report)
• History-Aware Processing: Compressed context management for ticker-specific analysis
• Options Trading Focus: 2-week expiration strategies with greeks validation
• Performance Tracking: Hypothetical P&L simulation using real market data
• Duplicate Prevention: SHA-256 hash-based deduplication system
• Real-time Updates: WebSocket-driven live dashboard with instant notifications

Architecture Overview

The system follows a microservices architecture with Docker containerization:

graph TD
    A[News Sources] --> B[Ingestion Service]
    B --> C[Deduplication Engine]
    C --> D[Pre-Pull Data Service]
    D --> E[AI Orchestrator]
    E --> F[Research Phase]
    F --> G[Report Generation]
    G --> H[WebSocket Push]
    H --> I[Live Dashboard]

    J[History Management] --> E
    K[Compression Engine] --> J
    L[Performance Tracker] --> G

AI Workflow Process

graph TD
    A[New Article Detected] --> B[Load Compressed History]
    B --> C[Pre-pull Market Data]
    C --> D[Research Phase - Max 5 Iterations]
    D --> E{Sufficient Data?}
    E -->|No| F[Tool Calls: get_alpaca_data, browse_page]
    F --> D
    E -->|Yes| G[complete_research Tool]
    G --> H[Report Phase - Populate Sections]
    H --> I[finalize_report Tool]
    I --> J[Push to Live Wire]
    J --> K[Queue for Hourly Aggregation]

Technical Implementation

• Frontend: React, TypeScript, Material-UI, Redux, Socket.io-client
• Backend: Node.js, Express, TypeScript, node-cron
• AI Integration: Deepseek via OpenAI SDK with MCP tool framework
• Data Storage: SQLite for history, Redis for caching
• APIs: Alpaca (SIP/OPRA feeds), Financial Modeling Prep
• Deployment: Docker Compose with local-only infrastructure