← Chat Archive

Here is a high-level plan for creating a new trading bot: **Step 1 — Define the Strategy** - Core logic (e.g. Renko VIDYA, 3-Candle pattern, momentum breakout) - Entry/exit rules, stop loss, target - Timeframe and instruments (NIFTY/BANKNIFTY futures, options, etc.) **Step 2 — Set Up the Bot Repository** - Create a new directory under `/home/ubuntu/bots/` (e.g. `/home/ubuntu/bots/<name>_bot`) - Copy or scaffold a `bot.py` (main execution loop), `strategy.py` (signal logic), `.env` (broker credentials) - Follow existing conventions from `claude_bot` or `lt_3candle_bot` for consistency **Step 3 — Broker Integration** - Choose broker: Shoonya (recommended) or Flattrade — both support historical data + WebSocket - Kotak is NOT suitable for new bots (no historical data support = no backtesting) - Configure `broker_api.py` and verify authentication flows **Step 4 — Backtesting** - Pull historical candle data via Shoonya `/EODChartData` or `/TPSeries` endpoints - Run strategy against 3-6 months of data to validate edge - Iterate on parameters if needed **Step 5 — Paper Trading** - Connect to live market feed (WebSocket) in paper mode - Track paper P&L, drawdown, signal hit rate **Step 6 — Telegram Alerts (Optional)** - Integrate `telegram_bot.py` for signal notifications - Follow the pattern from `claude_bot/telegram_bot.py` **Step 7 — Deploy as Systemd Service** - Create `/etc/systemd/system/<name>_bot.service` - Enable auto-restart on failure - Add log rotation --- Which broker do you want to use, and what strategy or timeframe are you thinking?