Bank Nifty Algo Trading: Zerodha API, Python Setup & SEBI Legal Framework

Manual Bank Nifty trading has two enemies: speed (you cannot react in milliseconds) and emotion (you override your own rules). Algorithmic trading eliminates both. This guide shows you how to set up a Bank Nifty options algo using Zerodha's Kite Connect API and Python — from zero to running your first automated trade.

SEBI Legal Framework for Algo Trading

SEBI's circular dated 16 March 2023 established rules for algo trading by retail investors:

• API-based orders are permitted: Retail traders can use broker APIs to place automated orders. This is explicitly legal.
• No exchange approval needed: Unlike institutional algos, retail API-based trading does not require NSE/BSE approval.
• Broker must provide API: Brokers like Zerodha (Kite Connect), Fyers (Fyers API), and Angel One (Smart API) offer official APIs.
• Rate limits apply: Zerodha allows 3 orders per second per user. Exceeding this triggers a temporary block.
• No co-location: Retail algo traders cannot use exchange co-location servers. Your latency will be 50-200ms — fine for strategies that are not latency-sensitive.

Zerodha Kite Connect API Setup

1. Subscribe to Kite Connect: Go to developers.kite.trade → Sign up → ₹2,000/month subscription. This gives you API access.
2. Create an app: In the Kite Connect dashboard, create a new app. Note your api_key and api_secret.
3. Generate access token: The login flow requires manual authentication once per day (Kite uses TOTP for security). After login, you receive an access_token valid for the entire trading session.
4. Install Python library: pip install kiteconnect

Python Environment

Required libraries: kiteconnect (Zerodha API), pandas (data manipulation), numpy (calculations), schedule (time-based triggers), logging (trade logging). Optional: scipy (Black-Scholes calculations), telegram-send (notifications).

The basic flow: authenticate → fetch Bank Nifty spot and option chain → apply strategy logic → place order → manage position → log everything.

Automating a Short Straddle Strategy

Example: automated short straddle with adjustment rules.

1. At 9:20 AM: Fetch Bank Nifty spot. Find ATM strike (nearest 100). Sell ATM CE + ATM PE. Log entry premiums.
2. Every 5 minutes: Check combined premium. If combined P&L exceeds -30% of collected premium, trigger SL exit.
3. If one leg doubles: Close both legs, re-sell straddle at new ATM. Maximum 2 adjustments.
4. At 3:15 PM: Close all positions regardless of P&L. No overnight hold.
5. Log: Entry prices, exit prices, adjustments, P&L, VIX at entry, PCR at entry. Save to CSV for backtesting.

Critical Risk Controls

• Daily loss limit: If total day P&L exceeds -₹10,000 (for ₹5L capital), stop all trading for the day. Hard-coded, not overridable.
• Position limit: Maximum 2 lots per strategy. Hard-coded.
• Kill switch: A manual kill switch (Telegram command or keyboard shortcut) that instantly closes all positions and stops the algo.
• Network monitoring: If the API connection drops for more than 30 seconds, close all positions via a backup manual order.
• No overnight positions: The algo must close all positions before 3:25 PM. Hard-coded deadline.
• Paper test first: Run the algo on Sensibull virtual trading for minimum 2 weeks before deploying real capital.

Alternative Platforms

Best free option: Angel One Smart API — no monthly charge, full F&O access, Python support. The trade-off is slightly slower execution compared to Zerodha.

Frequently Asked Questions