brokerkit-zerodha

BrokerKit's Zerodha adapter — wraps the official kiteconnect SDK behind BrokerKit's broker-agnostic interfaces (auth, instruments, orders, portfolio, market data, historical candles, streaming), plus two Zerodha-only extras: charges and GTT.

pip install brokerkit-core brokerkit-zerodha

Read this first: two things Zerodha does differently

1. There is no programmatic login. At all. Kite Connect has no TOTP, no password grant, nothing headless — verified against the SDK source (zero TOTP/2FA references anywhere in kiteconnect) and the official docs, which describe only the browser flow. You open Zerodha's login page, log in there, and the redirect carries a request_token you exchange for an access_token. That token expires at 6:00 AM IST the next day, so this is a once-a-day step.

renew_access_token exists in the SDK, but Zerodha only issues the refresh_token it needs to "certain approved platforms" — a personal app generally never receives one, so there is no unattended refresh.

2. The free plan has no market data. This is the inverse of Groww and Dhan:

So on a free app, market, historical and streaming fail with a permission error. That is your subscription, not a bug — the adapter's error message says so explicitly.

There is also no sandbox of any kind (unlike Upstox and Dhan), so order writes additionally need SEBI static-IP registration with no way around it.

Prerequisites

1. Create an app at developers.kite.trade — you get an api_key and api_secret, and you register a redirect URL (use a plain http://127.0.0.1:<port>/).
2. Static IP registration for order placement (SEBI rule). Reads work without it.
3. The ₹500/mo Connect plan, only if you need market data / historical / streaming.

Quick start

Mint a token once a day:

python -m brokerkit_zerodha.login_helper <api_key> <api_secret> http://127.0.0.1:5001/

This opens the browser and captures the redirect with a local server — nothing to copy-paste. (Port 5001, not 5000: macOS AirPlay Receiver squats on 5000 by default.)

Then reuse the printed token all day:

import asyncio
import os

from brokerkit import Exchange, Segment, create_broker


async def main():
    broker = await create_broker(
        "zerodha",
        api_key=os.environ["ZERODHA_API_KEY"],
        api_secret=os.environ["ZERODHA_API_SECRET"],
        access_token=os.environ["ZERODHA_ACCESS_TOKEN"],
    )

    instruments = await broker.instruments.fetch_instruments()
    reliance = next(
        i for i in instruments
        if i.symbol == "RELIANCE"
        and i.exchange == Exchange.NSE
        and i.segment == Segment.CASH
    )

    print(await broker.portfolio.holdings())
    await broker.close()


asyncio.run(main())

Omit access_token and pass redirect_uri= instead, and create() will run the browser login itself — right for an interactive first run, wrong for anything unattended.

Extras

broker.charges — pre-trade cost estimate

Implements the core ChargesProvider via Kite's virtual contract note.

from decimal import Decimal
from brokerkit.enums import Product, TransactionType

charges = await broker.charges.get_brokerage(
    reliance, quantity=10, product=Product.CNC,
    transaction_type=TransactionType.BUY, price=Decimal("1400"),
)
print(charges.total, charges.taxes.stt, charges.other_charges.sebi_turnover)

Works on the free plan.

broker.gtt — Good-Till-Triggered orders

A standing instruction that sits on Zerodha's servers (up to a year) and fires a real order when price crosses a trigger. No other broker in BrokerKit has this, so it is adapter-local with its own models rather than forced into OrderProvider — GTTs are a separate order book with their own endpoints and lifecycle, not an order type.

from decimal import Decimal
from brokerkit.enums import TransactionType
from brokerkit_zerodha import GttLeg

await broker.gtt.place(
    reliance,
    trigger_values=[Decimal("1500")],
    last_price=Decimal("1400"),          # current LTP; Kite validates against it
    legs=[GttLeg(transaction_type=TransactionType.SELL,
                 quantity=1, price=Decimal("1500"))],
)

for t in await broker.gtt.list_triggers():
    print(t.trigger_id, t.trading_symbol, t.status, t.trigger_values)

Two-leg (OCO — stop-loss and target together, whichever fires cancels the other) is supported with trigger_type="two-leg", two ascending trigger values and two legs.

Adapter notes

• Instrument master is already in rupees. Unlike Dhan, Angel One and Upstox, no paise division is needed — tick_size and strike come through as-is. The master is a public CSV needing no auth (122,526 rows → 82,715 normalized instruments).
• Index rows carry instrument_type "EQ". Kite's master has only four type values (EQ/FUT/CE/PE) and no index type; indices are identifiable only by segment == "INDICES". Trusting instrument_type alone would classify all 220 indices as tradeable equities. The adapter maps them to InstrumentType.IDX correctly.
• No option-chain endpoint and no greeks endpoint. get_option_chain is assembled from the master plus batched quotes. OptionContract.greeks is always None here — Kite has nothing to merge. Use the Fyers or Upstox adapter if you need greeks.
• Instrument.isin is always None — Kite's master has no ISIN column. Join against other brokers on exchange_token (RELIANCE is 2885 on Groww, Fyers, Dhan and Zerodha alike).
• Streaming runs on the Twisted global reactor. KiteTicker.stop() calls reactor.stop(), which is process-wide and irreversible, so this adapter's close() deliberately closes only the socket. "full" mode is used rather than "quote" because the quote packet carries no timestamp field at all.
• Tick.minute_ohlc is always None — Kite's feed has no server-computed minute candle. Aggregate ticks yourself.

Verification status

Live-verified against a real account: auth, instruments, portfolio, orders (read), and charges end to end.

Blocked by account state rather than by code: market / historical / streaming / option chain need the paid Connect plan; order and GTT writes need SEBI static IP, with no sandbox available to sidestep it.

License

MIT © 2026 Aditya Vishwakarma