FOREX AI Trading Bot

An intelligent FOREX trading bot powered by machine learning and ChatGPT strategic analysis.

Features

• 🤖 Local ML Trading Signals: Random Forest classifier for real-time BUY/SELL/HOLD predictions
• 🧠 ChatGPT Strategic Analysis: Daily market regime detection and performance insights
• 📊 Real-time Dashboard: Web interface with TradingView-style charts
• 📱 Telegram Bot: Mobile notifications and trade approvals
• 🔬 Backtesting Engine: Validate strategies on historical data
• 📈 Paper Trading: Safe testing with OANDA practice account
• ☁️ AWS Deployment: Scalable cloud infrastructure

Architecture

Three-Tier Hybrid Approach:

1. Local ML (Fast Layer): Random Forest for real-time signals (every 5 min)
2. ChatGPT (Smart Layer): Strategic analysis 1-3x per day
3. Human Oversight (Safety Layer): Manual approval via Telegram

For detailed system architecture, see docs/ARCHITECTURE.md.

For architectural decisions and rationale, see docs/DECISIONS.md.

Tech Stack

• Backend: Python 3.11+, FastAPI, Celery, SQLAlchemy
• ML/AI: scikit-learn, TA-Lib, OpenAI API (GPT-4o)
• Frontend: React, TypeScript, TradingView Lightweight Charts
• Database: PostgreSQL + TimescaleDB (time-series optimized)
• Cache: Redis
• Infrastructure: AWS (EC2, RDS, S3), Terraform
• Trading: OANDA v20 API

Quick Start

Prerequisites

• Python 3.11+
• Poetry (dependency management)
• Docker & Docker Compose
• OANDA practice account
• OpenAI API key

Installation

1. Clone the repository:

git clone <your-repo-url>
cd forex-ai-trading-bot

2. Copy environment template:

cp .env.example .env

3. Edit .env and add your API keys:

• OANDA_API_KEY
• OPENAI_API_KEY
• TELEGRAM_BOT_TOKEN (optional)

4. Start local development environment:

cd infrastructure/docker
docker-compose up -d

5. Install Python dependencies:

cd backend
poetry install
poetry shell

6. Run database migrations:

alembic upgrade head

7. Download historical data:

python scripts/download_historical_data.py

8. Start the services:

# Terminal 1: API server
uvicorn backend.api.main:app --reload

# Terminal 2: Celery worker
celery -A backend.strategy_engine.celery_app worker --loglevel=info

# Terminal 3: Celery beat (scheduler)
celery -A backend.strategy_engine.celery_app beat --loglevel=info

Project Structure

forex-ai-trading-bot/
├── backend/
│   ├── shared/              # Shared utilities (config, database, logger)
│   ├── data_ingestion/      # OANDA API client, data streaming
│   ├── strategy_engine/     # ML models, signals, LLM integration
│   ├── execution/           # Order execution, position management
│   ├── api/                 # FastAPI REST API and WebSocket
│   ├── telegram_bot/        # Telegram bot for notifications
│   └── backtesting/         # Backtesting engine
├── frontend/                # React dashboard
├── infrastructure/          # Docker, Terraform
├── docs/                    # Documentation
│   ├── ARCHITECTURE.md      # System design and component details
│   ├── DECISIONS.md         # Architecture Decision Records (ADRs)
│   └── (future) SPEC.md     # Technical specifications
├── scripts/                 # Utility scripts
└── notebooks/               # Jupyter notebooks for analysis

Configuration

Key configuration in .env:

• TRADING_PAIRS: Comma-separated list (default: EUR/USD,GBP/USD,USD/JPY)
• PAPER_TRADING: Always true for safety
• ML_CONFIDENCE_THRESHOLD: Minimum confidence for trades (default: 0.65)
• LLM_MAX_CALLS_PER_DAY: ChatGPT API call limit (default: 10)

Trading Strategy

Trend-Following with ML Confirmation:

• Entry when EMA(12) > EMA(26), RSI 40-70, MACD bullish, ML confidence > 0.65
• Exit: Stop Loss at 1.5×ATR, Take Profit at 2.5×ATR
• Risk: 1% per trade, max 3 concurrent positions

ChatGPT Strategic Overlay:

• Morning: Market regime analysis (trending vs ranging)
• Evening: Performance review
• On news events: Sentiment analysis
• Weekly: Correlation analysis

Development

Running Tests

pytest
pytest --cov=backend tests/

Code Formatting

black backend/
isort backend/
flake8 backend/

Database Migrations

# Create migration
alembic revision --autogenerate -m "description"

# Apply migrations
alembic upgrade head

Deployment

See PLAN.md for detailed deployment guide.

AWS Free Tier (Year 1):

• EC2 t2.micro/t3.micro
• RDS db.t2.micro/t3.micro
• S3, CloudWatch (free tier)
• Cost: $0-5/month + $20 ChatGPT Plus

Monitoring

• CloudWatch: System metrics, logs, alarms
• Dashboard: Real-time P&L, positions, signals
• Telegram: Trade notifications, daily summaries

Safety Features

• Paper trading only (initially)
• Daily loss circuit breaker (3% of capital)
• Position limits (max 3 concurrent)
• ChatGPT never directly executes trades
• All trades logged and auditable

Roadmap

• Phase 1: Foundation (Data pipeline, API)
• Phase 2: ML & ChatGPT Integration
• Phase 3: Execution & Backtesting
• Phase 4: User Interfaces (Dashboard, Telegram)
• Phase 5: AWS Deployment
• Phase 6: Optimizations (LSTM, more strategies)

License

MIT License - See LICENSE file

Disclaimer

⚠️ This is for educational and paper trading purposes only. Trading financial instruments carries significant risk. Never trade with money you cannot afford to lose. Past performance does not guarantee future results.

Documentation

• PLAN.md: 15-week implementation roadmap
• docs/ARCHITECTURE.md: Detailed system architecture and component design
• docs/DECISIONS.md: Architecture Decision Records (ADRs) explaining key technical choices

Support

For questions or issues, please open a GitHub issue or refer to the documentation above.