Trading Order

Autonomous multi-strategy algorithmic trading system for Indian markets (NSE), built with Node.js.
It combines market data, curated financial news sentiment, and Anthropic Claude-based decision logic with strict risk controls and paper-trading analytics.

Features

• Multi-strategy decisioning for F&O intraday and equity swing logic.
• Claude-powered macro bias, sentiment scoring, and trade recommendations.
• Curated financial news ingestion (CNBC-TV18, ET, Mint, Moneycontrol, NSE, and more via domain filters).
• Strict risk framework: per-trade allocation cap, mandatory SL/TP, and daily drawdown halt.
• IST market scheduler (pre-market, live loop, square-off, post-market analysis).
• Paper-trading mode with deep JSONL telemetry and daily analytics reports.
• Optional Redis-backed state store (falls back to in-memory).

Architecture

• src/app.js: Bootstrap and dependency wiring.
• src/scheduler.js: Indian market-hours cron orchestration.
• src/tradingEngine.js: Core execution loop, paper/live handling, risk checks, logging.
• src/aiEngine.js: Anthropic client and strict JSON response parsing.
• src/growwClient.js: Groww REST integration with exponential backoff.
• src/newsClient.js: Multi-provider + curated-domain financial news collector.
• src/riskManager.js: Position sizing, SL/TP enforcement, drawdown checks.
• src/store.js: In-memory/Redis state for positions and daily trading context.

Requirements

• Node.js >=18.18.0 (recommended)
• npm
• API keys for Anthropic + News provider(s)
• Groww credentials for live market data / live execution

Note: On older Node versions, npm run dev may fail due to --watch. Use npm start.

Setup

1. Install dependencies:

npm install

2. Copy and configure environment variables:

cp .env.example .env

3. Update required keys in .env:

• ANTHROPIC_API_KEY
• NEWSAPI_KEY (for curated/newsapi mode)
• Groww keys/tokens for market data or live execution

Run

Production mode:

npm start

Development mode (Node 18+):

npm run dev

Paper Trading (Safe Testing)

Set in .env:

ENABLE_LIVE_TRADING=false
PAPER_TRADING=true
PAPER_TRADING_SLIPPAGE_PCT=0.05

Then run:

npm start

Paper mode behavior:

• Simulates order placement and square-off.
• Tracks positions and realized PnL in internal state.
• Produces structured logs for analysis.

Logging & Analytics Outputs

• paper_trades.jsonl: Paper order + square-off events.
• deep_analysis.jsonl: Per-cycle telemetry, API timings, decisions, rejections, risk events.
• paper_daily_report.jsonl: End-of-day summary metrics (PnL, win rate, profit factor, etc.).

These files are ignored by git via .gitignore.

Scheduler Windows (IST)

• 08:30 Pre-market: macro scan + news analysis.
• 09:15 to 15:30 Live cycle loop.
• 15:15 Intraday/F&O forced square-off.
• 15:30 to 17:00 Post-market reporting windows.

Risk Controls

• Max trading capital from MAX_TRADING_AMOUNT.
• Single high-risk F&O trade allocation capped at 10% (or lower if configured).
• Mandatory stop-loss and target enforcement on each trade.
• Global halt when daily drawdown breaches DAILY_MAX_DRAWDOWN_PCT.

Important Notes

• Groww endpoints in the client are scaffold-style and may need alignment with your exact broker API contract.
• This project is for educational/research use. Validate thoroughly in paper mode before enabling live trading.
• You are responsible for regulatory and broker compliance.