MIRA - AI-Powered Prediction Market Terminal

Real-time AI-powered prediction market interface with live trading analytics and neural network visualizations.

Features

AI Agent System

• Multiple AI models for diverse trading strategies
• Real-time decision making and market analysis
• Performance tracking and analytics

Prediction Market Integration

• Polymarket API integration
• Real-time market data
• Category filtering and search

Trading Interface

• Interactive bubble visualization
• Zoom and pan controls
• Real-time position tracking
• Trade execution on Solana

Agent Builder

• 3-step wizard interface
• Strategy configuration
• Risk management settings
• Live testing and deployment

AI Agent Decision-Making Flowchart

The following flowchart illustrates how AI agents analyze markets and make trading decisions:

flowchart TD
    A[Agent Cycle Starts] --> B[Fetch Markets & News<br/>from Polymarket API]
    B --> C[Check Cache<br/>for Recent Trades]
    C -->|Cache Hit| D[Return Cached Trades]
    C -->|Cache Miss| E[Filter Candidate Markets<br/>Volume, Liquidity, Category]
    E --> F[Score Each Market<br/>Volume, Liquidity, Price, News, Probability]
    F --> G[Sort by Score<br/>Take Top Markets]
    G --> H[For Each Top Market]
    H --> I{Score >= 8?}
    I -->|No| J[Skip Market]
    I -->|Yes| K[Search Web<br/>for Market Info]
    K --> L{AI API<br/>Configured?}
    L -->|No| M[Use Deterministic<br/>Fallback Logic]
    L -->|Yes| N{Check AI Cache<br/>5min TTL}
    N -->|Cache Hit| O[Use Cached<br/>AI Decision]
    N -->|Cache Miss| P[Call AI API<br/>GPT-5, Claude, GROK, etc.]
    P --> Q[Parse AI Response<br/>Side, Confidence, Reasoning]
    Q --> R[Cache AI Decision]
    R --> S[Apply Personality Rules<br/>Agent-Specific Adjustments]
    O --> S
    M --> S
    S --> T[Calculate Position Size<br/>Based on Risk, Confidence, Score]
    T --> U[Generate Trade<br/>Side, Confidence, Investment]
    U --> V{Reached<br/>Max Trades?}
    V -->|No| H
    V -->|Yes| W[Cache All Trades<br/>30s TTL]
    W --> X[Return Trades]
    J --> V
    D --> X

Loading

Flow Explanation

1. Agent Cycle Starts: Each agent runs its trading cycle independently
2. Fetch Markets & News: Agent fetches all available markets from Polymarket API and latest news articles
3. Check Cache: System checks if trades were recently generated (30s cache) to avoid redundant AI calls
4. Filter Candidate Markets: Markets are filtered by:
   • Minimum volume (agent-specific: $50k-$200k)
   • Minimum liquidity (agent-specific: $10k-$50k)
   • Category preferences (if agent has focus categories)
5. Score Each Market: Markets are scored using weighted components:
   • Volume score (trading activity)
   • Liquidity score (market depth)
   • Price movement score (24h change)
   • News score (recent relevant news)
   • Probability score (current market probability)
6. Sort by Score: Markets are sorted by total weighted score, top markets selected
7. For Each Top Market: Agent analyzes each high-scoring market
8. Score Threshold Check: Markets with score < 8 are skipped
9. Search Web: Agent searches the web for market-specific information to inform decision
10. AI Decision Path:
    • If AI API configured: Check AI cache (5min TTL), if miss call AI API (GPT-5, Claude, GROK, Gemini, DeepSeek, or Qwen)
    • If AI API fails/unconfigured: Use deterministic fallback based on market data
11. Parse AI Response: Extract side (YES/NO), confidence (0-1), and reasoning from AI
12. Apply Personality Rules: Agent-specific adjustments based on personality (risk tolerance, category preferences, etc.)
13. Calculate Position Size: Investment amount calculated from:
    • Base risk budget (LOW: $80, MEDIUM: $150, HIGH: $250)
    • Confidence multiplier (0.5x to 2.5x)
    • Market score multiplier (0.4x to 2.0x)
    • Personality adjustments
14. Generate Trade: Create trade object with side, confidence, investment, reasoning
15. Check Max Trades: Continue until agent reaches max trades limit (varies by agent: 3-10 trades)
16. Cache All Trades: Generated trades are cached for 30 seconds to avoid regeneration
17. Return Trades: Agent returns all generated trades for display and execution

Tech Stack

• Frontend: React + TypeScript + Vite
• Backend: Express.js + Node.js
• UI: shadcn/ui + Tailwind CSS
• Blockchain: Solana Web3.js
• Deployment: Railway

Development

Prerequisites

• Node.js 18+
• npm

Installation

npm install

Run Development Server

Start both frontend and backend:

npm run dev:all

Or separately:

# Terminal 1: Backend
npm run server

# Terminal 2: Frontend
npm run dev

The application will be available at:

• Frontend: http://localhost:3000
• Backend: http://localhost:3002

Build

Build for production:

npm run build

Environment Variables

Backend (Railway/Server)

Required:

• PORT - Server port (auto-set by Railway)
• NODE_ENV - Environment (production/development)

Optional:

• REDIS_URL - Redis connection URL for session storage (highly recommended for production)
  • How to get: Add Redis addon in Railway dashboard → Your Service → "+ New" → "Database" → "Add Redis"
  • Railway will automatically set this variable when you add the Redis addon
  • Without Redis, sessions are stored in memory (lost on restart, won't work across multiple instances)
• ALLOWED_ORIGINS - Comma-separated list of allowed CORS origins
• CSRF_SECRET - Secret for CSRF token generation (recommended for production)
• SESSION_SECRET - Secret for session management (required for Google OAuth, recommended for production)
• GOOGLE_CLIENT_ID - Google OAuth Client ID (required for Google login)
• GOOGLE_CLIENT_SECRET - Google OAuth Client Secret (required for Google login)
• GOOGLE_CALLBACK_URL - Google OAuth callback URL (defaults to https://mira.tech/api/auth/google/callback in production)
• NEWS_API_KEY - NewsAPI.org API key
• NEWSDATA_API_KEY - NewsData.io API key
• GNEWS_API_KEY - GNews API key
• POLYMARKET_API_KEY - Polymarket API key
• POLYMARKET_SECRET - Polymarket API secret
• POLYMARKET_PASSPHRASE - Polymarket API passphrase
• SMTP_HOST - SMTP server host
• SMTP_PORT - SMTP server port
• SMTP_USER - SMTP username
• SMTP_PASS - SMTP password
• NOTIFICATION_EMAIL - Email for waitlist notifications

Frontend

Optional:

• VITE_API_BASE_URL - Custom API base URL (defaults to relative URLs in production)

Project Structure

aura-predict/
├── src/
│   ├── components/     # React components
│   │   ├── ui/         # shadcn/ui components
│   │   └── ...         # Feature components
│   ├── pages/          # Page components
│   ├── hooks/          # Custom React hooks
│   ├── lib/            # Utility libraries and API clients
│   └── types/          # TypeScript type definitions
├── server/
│   ├── services/       # Business logic services
│   ├── utils/          # Utility functions
│   └── index.js        # Express server
├── public/             # Static assets
└── dist/               # Production build output

Google OAuth Setup

To enable Google login:

1. Create Google OAuth Credentials:

   • Go to Google Cloud Console
   • Create a new project or select an existing one
   • Enable the Google+ API
   • Go to "Credentials" → "Create Credentials" → "OAuth client ID"
   • Choose "Web application"
   • Add authorized redirect URIs:
     • Production: https://mira.tech/api/auth/google/callback
     • Development: http://localhost:3002/api/auth/google/callback

2. Set Environment Variables:

   GOOGLE_CLIENT_ID=your-client-id.apps.googleusercontent.com
   GOOGLE_CLIENT_SECRET=your-client-secret
   SESSION_SECRET=your-session-secret  # Generate with: openssl rand -base64 32

3. Deploy:

   • Add these variables to your Railway project settings
   • The OAuth flow will work automatically once configured

Security

The application includes comprehensive security features:

• CSRF protection
• Rate limiting
• Input validation
• Security headers (Helmet.js)
• CORS configuration
• Error sanitization
• Google OAuth authentication
• Secure session management

License

This project is private and proprietary.

Contact

For inquiries about MIRA, please contact dev@mira.tech

---

Built for the prediction market community