Crypto Trading AI Assistant

Educational AI-powered cryptocurrency trading assistant with market simulation, technical analysis, and risk management.

⚠️ WARNING: This is an educational and analytical tool, NOT for real automated trading. Always make your own informed decisions. Trading cryptocurrencies carries significant risk.

Features

• 📊 Technical Indicators: RSI, MACD, EMA, SMA, Bollinger Bands
• 🎯 Trading Strategies: Scalping, Intraday, Swing, Position strategies
• 🤖 AI Trading Bot: Multi-layered simulation bot with market analysis
• 📉 Market Simulator: Generate artificial candles with trends, volatility, and shocks
• ⚖️ Risk Management: Liquidation price calculation, leverage risk assessment
• 💼 Portfolio Analysis: Multi-exchange portfolio aggregation and risk metrics
• 📰 News Aggregation: Crypto news and sentiment analysis
• 🤖 OpenAI Integration: LLM-powered market analysis (optional)
• 🔒 Read-Only Exchange Integration: KuCoin and Gate.io via ccxt
• 📱 Telegram Bot: Basic bot framework (optional)

Installation

npm install

Configuration

Copy .env.example to .env and configure:

cp .env.example .env

Environment Variables

Below is a complete example of the .env file (all keys are optional, but exchange keys enable live data):

# LLM Providers
OPENAI_API_KEY=your_openai_api_key_here
DEEPSEEK_API_KEY=your_deepseek_api_key_here

# KuCoin (read-only)
KUCOIN_API_KEY=your_kucoin_api_key
KUCOIN_API_SECRET=your_kucoin_api_secret
# Passphrase is created when you generate the API key (it is NOT the trading password)
KUCOIN_API_PASSPHRASE=your_kucoin_api_passphrase

# Gate.io (read-only)
GATE_API_KEY=your_gate_api_key
GATE_API_SECRET=your_gate_api_secret

# Optional integrations
TELEGRAM_BOT_TOKEN=your_telegram_bot_token

Notes:

• Exchange API keys are used in read-only mode; no live trading is executed.
• KuCoin passphrase is set by you during API key creation and is required for request signatures.

CLI Commands

1. Generate Market Simulation

npm start simulate -- --symbol TONUSDT --timeframe 15m --candles 500 --initial-price 2.5

Options:

• --volatility <number>: Market volatility (default: 0.02)
• --trend-strength <number>: Trend strength 0-1 (default: 0.3)
• --shock-probability <number>: Probability of price shocks (default: 0.05)

2. Run Trading Bot Simulation

npm start trade-sim -- --symbol TONUSDT --timeframe 15m --candles 1000 --initial-price 2.5

Options:

• --initial-balance <number>: Starting balance in USD (default: 10000)
• --max-leverage <number>: Maximum leverage (default: 5)
• --mmr <number>: Maintenance margin rate (default: 0.005)
• --aggressiveness <number>: Trading aggressiveness multiplier (0.5-2.0, default: 1)
• --report <path>: Inject CSV/TSV/Excel performance reports to bias risk when historical PnL is negative.
• --save-chart / --export-report pdf|json|csv / --no-llm for saving charts, exporting reports, and skipping LLM output.

3. Analyze Real Trading Pair

npm start analyze-pair -- --exchange kucoin --symbol TON/USDT --timeframe 15m --limit 200

Requires exchange API keys configured in .env.

• --since YYYY-MM-DD or --months N fetch multi-month OHLCV via paginated requests; --no-cache bypasses cached candles.
• OHLCV loading defaults to --ohlcv-source auto (cache + fetch missing), and CCXT rate limiting is enabled by default (override with --no-rate-limit).
• Indicators now include OBV/VWAP/funding rate and are shown in the console output.

4. Analyze Portfolio

npm start analyze-portfolio

Aggregates balances from configured exchanges.

5. Analyze Crypto News

npm start analyze-news -- --symbol TON

Quick recipes

• посмотреть режим

npm start -- analyze-regime --exchange gate --symbol RAVE/USDT --since 2024-01-01 --ohlcv-source auto

Auto mode reads cached OHLCV first and fetches any missing segments; add --verbose to see detailed cache/fetch logs (or --quiet to suppress them).

• бэктест grid

npm start -- backtest-grid --symbol RAVE/USDT --since 2024-01-01 --mode spot

• сравнение с ledger

npm start -- compare ./data/ledger.csv --symbol RAVE/USDT

• PROMO MODE

npm start -- backtest-grid --profile promo --symbol RAVE/USDT --since 2024-01-01 --mode trailing --save-report

По умолчанию stop по MA30 считается на 15m свечах (используйте --ma-timeframe native, чтобы оставить нативный таймфрейм).

Report Analysis

Supply CSV/TSV/Excel reports containing day, spent, voucherIncome, and profit columns. The assistant aggregates totals and adjusts strategy confidence when losses dominate.

LLM Integration

The LLM layer prefers DeepSeek when DEEPSEEK_API_KEY is present (base URL https://api.deepseek.com/v1, default model deepseek-chat), otherwise it uses OpenAI via OPENAI_API_KEY with LLM_MODEL set. Use --no-llm to disable prompts. Outputs are scenario-based (conservative/moderate/aggressive) and avoid direct buy/sell instructions.

Historical Data & Funding

analyze-pair supports long lookbacks via --since or --months, invoking paginated fetchFullOHLCV. For perpetual pairs on sub-daily timeframes, funding rates are fetched and shown alongside OBV and VWAP.

New Indicators & Strategies

OBV, VWAP, funding-rate-aware scoring, and volume/funding/multi-timeframe strategies are available. Strategy weights are combined to derive the aggregate signal.

KuCoin API passphrase is the user-defined phrase set during API key creation (not the trading password).

Architecture

src/
├── core/           # Types, interfaces, utilities
├── indicators/     # Technical indicators (RSI, MACD, EMA, etc.)
├── strategies/     # Trading strategies (scalping, intraday, swing, position)
├── simulation/     # Market simulator, order execution, AI trader bot
├── real/           # Exchange clients, portfolio analyzer, risk math
├── news/           # News aggregation and sentiment analysis
├── llm/            # OpenAI integration and prompts
└── ui/             # CLI interface and Telegram bot

Multi-Layered AI Trading Bot

The simulation bot uses a sophisticated multi-layer architecture:

1. Market Layer: Fetches market state and price data
2. Indicator Layer: Computes technical indicators
3. Strategy Layer: Runs multiple strategies and combines signals
4. Risk Layer: Assesses liquidation risk and leverage
5. LLM Layer (Optional): AI-powered market analysis
6. Decision Layer: Makes final trading decisions
7. Execution Layer: Simulates order execution and PnL

Risk Disclaimer

• This tool is for educational purposes only
• NOT financial advice - always do your own research
• Cryptocurrency trading carries significant risk
• Leverage trading can result in total loss of capital
• Past performance does not guarantee future results
• The developers are not responsible for any financial losses

Testing

npm test

License

MIT

Contributing

Contributions welcome! This is an educational project focused on learning trading strategies, risk management, and AI integration.

TODO for Users

• Add more technical indicators (ATR, Stochastic, etc.)
• Implement additional trading strategies
• Integrate more news sources
• Add web interface for visualization
• Enhance backtesting with historical data
• Implement portfolio optimization algorithms