Kraken

Kraken offers spot and derivatives trading across a wide range of digital assets. This integration connects to Kraken Pro and supports live market data ingest and order execution for Kraken Spot and Kraken Derivatives (Futures).

Overview

This adapter is implemented in Rust with Python bindings for ease of use in Python-based workflows. It does not require external Kraken client libraries; the core components are compiled as a static library and linked automatically during the build.

This guide assumes a trader is setting up for both live market data feeds and trade execution. The Kraken adapter includes multiple components, which can be used together or separately depending on the use case.

• KrakenSpotRawHttpClient and KrakenFuturesRawHttpClient: Low-level HTTP API connectivity.
• KrakenSpotHttpClient and KrakenFuturesHttpClient: Higher-level HTTP clients with instrument caching and reconciliation support.
• KrakenInstrumentProvider: Instrument parsing and loading functionality.
• KrakenDataClient: Market data feed manager.
• KrakenExecutionClient: Account management and trade execution gateway.
• KrakenLiveDataClientFactory: Factory for Kraken data clients (used by the trading node builder).
• KrakenLiveExecClientFactory: Factory for Kraken execution clients (used by the trading node builder).

Examples

You can find live example scripts here.

Kraken documentation

Kraken provides detailed documentation for users:

• Kraken API Documentation
• Kraken Spot REST API
• Kraken Futures REST API

Refer to the Kraken documentation in conjunction with this NautilusTrader integration guide.

Products

Kraken supports two primary product categories:

Bar streaming

Supported intervals

The Kraken adapter supports real-time bar (OHLC) streaming for Spot markets via WebSocket. The following intervals are available:

Bar emission latency

Kraken's WebSocket OHLC channel pushes updates for the current (incomplete) bar on every trade. Unlike some exchanges (e.g., Binance), Kraken does not provide an "is_closed" indicator to signal when a bar is complete.

To avoid emitting partial/incomplete bars, the adapter buffers the current bar and only emits it when the next bar period begins (i.e., when a message with a new interval_begin timestamp arrives). This means:

• Bars are emitted with a delay of up to one bar period.
• For 1-minute bars, the maximum delay is ~1 minute.
• The emitted bar data is complete and final.

We chose this approach over timer-based emission because:

• Timer-based emission could miss the final update before the bar closes.
• Kraken's updates are not guaranteed to arrive at exact interval boundaries.
• Buffering preserves data integrity at the cost of latency.

Symbology

Bitcoin symbol format (BTC vs XBT)

Kraken uses different Bitcoin symbol conventions across their APIs:

Spot markets

NautilusTrader uses ISO 4217-A3 format for Kraken Spot instrument symbols, which provides a standardized representation across exchanges. The adapter handles translation to Kraken's native format internally.

Instrument ID format:

InstrumentId.from_str("BTC/USD.KRAKEN")   # Spot BTC/USD
InstrumentId.from_str("ETH/USD.KRAKEN")   # Spot ETH/USD
InstrumentId.from_str("SOL/USD.KRAKEN")   # Spot SOL/USD
InstrumentId.from_str("BTC/USDT.KRAKEN")  # Spot BTC/USDT
InstrumentId.from_str("ETH/BTC.KRAKEN")   # Spot ETH/BTC (normalized from ETH/XBT)

Futures markets

Kraken Futures instruments use a specific naming convention with prefixes:

• PI_ - Perpetual Inverse contracts (e.g., PI_XBTUSD)
• PF_ - Perpetual Fixed-margin contracts (e.g., PF_XBTUSD)
• FI_ - Fixed maturity Inverse contracts (e.g., FI_XBTUSD_230929)
• FF_ - Flex futures contracts

Instrument ID format:

InstrumentId.from_str("PI_XBTUSD.KRAKEN")  # Perpetual inverse BTC
InstrumentId.from_str("PI_ETHUSD.KRAKEN")  # Perpetual inverse ETH
InstrumentId.from_str("PF_XBTUSD.KRAKEN")  # Perpetual fixed-margin BTC

Data capability

Subscriptions (real-time)

Requests (historical)

Orders capability

Order types

Time in force

Execution instructions

Trigger types

Conditional orders (stop, take-profit, trailing stop) support a trigger price reference on Spot:

Batch operations

Position management

Order querying

Contingent orders

Order routing (Spot)

The Rust Spot execution client routes submit_order, modify_order, cancel_order, and submit_order_list through Kraken's authenticated WebSocket v2 trade channel by default, with REST as the fallback. The Python live KrakenExecutionClient currently routes all orders via REST regardless of the knobs below; WebSocket trade routing is only active when the Rust execution client (KrakenSpotExecutionClient) is in use, either via the Rust factory or by constructing the pyo3-exposed nautilus_trader.core.nautilus_pyo3.kraken.KrakenExecClientConfig directly.

Order shapes routed via REST

Some Spot order shapes always route via REST. They split into two categories: shapes Kraken's WS v2 API does not support at all, and shapes the WS API supports but this adapter does not yet encode.

Kraken WS v2 limitation:

Not yet encoded by this adapter (follow‑up work, currently REST):

The per-call params={"use_ws_trade": False} override forces a single command through REST regardless of the configured default. Set it on SubmitOrder, ModifyOrder, CancelOrder, or SubmitOrderList.

WebSocket request timeout

When a WebSocket round-trip exceeds ws_request_timeout_secs (default 5) the dispatcher synthesises a local rejection event stamped with the timeout-fire timestamp:

• Submit / batch_add: OrderRejected (one per leg for batches). The dispatcher then sends a best-effort compensating cancel_order over the same WebSocket so a delayed venue acceptance is not left as an orphan order.
• Modify: OrderModifyRejected.
• Cancel: OrderCancelRejected.

The timeout does not trigger an automatic REST retry; strategies must resubmit if they want to try again. If the venue actually accepted the order and the compensating cancel does not land, the live execution reconciliation engine (open_check_interval_secs) is the recovery path.

Rust-side configuration knobs

The Rust KrakenExecClientConfig (and its pyo3 wrapper) exposes:

These are not exposed on the Python live KrakenExecClientConfig because the Python live execution client does not yet honour them.

Reconciliation

The Kraken adapter provides reconciliation capabilities for both Spot and Futures markets, allowing traders to synchronize their local state with the exchange state at startup or during operation.

Spot reconciliation

Order status reports:

• Open orders: Fetches all currently active orders.
• Closed orders: Fetches historical orders with pagination support.
• Time-bounded queries: Supports filtering by start/end timestamps.

Fill reports:

• Trade history: Fetches execution history with pagination.
• Time-bounded queries: Supports filtering by start/end timestamps.
• All fill types: Market, limit, and conditional order fills.

Margin position reports (when spot_account_type=Margin):

• Open positions: Fetched from POST /0/private/OpenPositions and aggregated by (pair, side) into PositionStatusReport entries.
• Synthetic FLAT cleanup: If the local cache has an open spot margin position that no longer appears on the venue (Kraken omits closed positions from OpenPositions), the adapter emits a synthetic FLAT report on the next position-check tick so the engine reconciles to closed.
• Margin balances: POST /0/private/TradeBalance is called alongside the account-state refresh; used margin populates MarginBalance.initial, remaining metrics flow into AccountState.info (see Spot margin trading).

Futures reconciliation

Order status reports:

• Open orders: Fetches all currently active futures orders.
• Historical orders: Fetches closed and filled orders when open_only=False.
• Order events: Full order lifecycle history via /api/history/v2/orders endpoint.

Fill reports:

• Fill history: Fetches all execution reports.
• Time filtering: Client-side filtering by start/end timestamps (parses RFC3339 timestamps).
• All fill types: Maker and taker fills with fee information.

Position status reports:

• Open positions: Fetches all active futures positions.
• Real-time data: Includes unrealized funding, average price, and position size.

Spot position reports (cash mode)

In cash mode, the Kraken adapter can optionally report wallet balances as position status reports for spot instruments. This feature is disabled by default and must be explicitly enabled via configuration. Margin-mode accounts should leave it disabled and rely on OpenPositions instead (see Spot margin trading).

How it works:

• When enabled, wallet balances are converted to PositionStatusReport objects.
• Positive balances are reported as LONG positions.
• Only instruments matching the configured quote currency are reported (default: USDT).
• This prevents duplicate reports when the same asset is available with multiple quote currencies (e.g., BTC/USD, BTC/USDT, BTC/EUR).

Configuration:

exec_clients={
    KRAKEN: {
        "use_spot_position_reports": True,
        "spot_positions_quote_currency": "USDT",  # Default
    },
}

Spot margin trading

Kraken Spot supports leveraged trading on selected pairs. Per-pair availability and the valid leverage tiers are advertised by Kraken on the instruments endpoint as AssetPairInfo.leverage_buy and leverage_sell; the adapter caches these at instrument-load time and validates the requested tier before order submission. Margin trading is enabled per-execution-client via spot_account_type, with per-order leverage params.

Configuration

from nautilus_trader.adapters.kraken import KrakenExecClientConfig
from nautilus_trader.model.enums import AccountType

exec_clients = {
    KRAKEN: KrakenExecClientConfig(
        spot_account_type=AccountType.MARGIN,
        default_leverage=3,             # optional config-level default
        margin_balance_asset="ZGBP",    # optional summary-display asset
    ),
}

margin_balance_asset controls only the denomination of the account-summary metrics returned by Kraken's TradeBalance endpoint (equity, free margin, used margin, etc.). Per-position figures from OpenPositions are always in the traded pair's quote currency.

Per-order leverage

Override the configured default on a single order via params:

order = strategy.order_factory.limit(
    instrument_id=BTC_USD,
    order_side=OrderSide.BUY,
    quantity=Quantity.from_str("0.01"),
    price=Price.from_str("50000.00"),
    params={"leverage": 5},
)

The adapter validates the requested tier against AssetPairInfo.leverage_buy / leverage_sell for the pair before submitting; an invalid tier produces an OrderDenied event and never hits the venue.

Reduce-only

Margin orders can carry reduce_only=True; Kraken rejects the order if no matching position exists. Cash orders ignore the flag.

Account state

When spot_account_type=Margin, the adapter calls Kraken's TradeBalance endpoint and surfaces the result in two places:

• MarginBalance.initial: used margin (m).
• AccountState.info dict: full TradeBalance snapshot:
  • equity: net equity
  • free_margin: equity minus used margin
  • unrealized_pnl: open-position P&L
  • margin_level: equity / used margin (%) when positions are open
  • trade_balance: collateral on deposit
  • equivalent_balance: combined-currency wallet equivalent
  • cost_basis, valuation, unexecuted_value, used_margin: raw TradeBalance fields
  • asset: resolved denominating asset (e.g. USD, GBP)

A single INFO log line is emitted on every account state refresh:

Margin metrics: equity=1234.56 GBP, free_margin=1100.00, unrealized_pnl=12.34

Strategies read the values via account_state.info["equity"], etc.

Position reconciliation

Open spot margin positions are surfaced via POST /0/private/OpenPositions on each position_check_interval_secs tick. Closed positions on the venue that still appear open in the local cache are reconciled to FLAT on the next sweep. This path is independent of use_spot_position_reports (which is wallet-derived, cash-mode-only).

Funding rates

The adapter receives funding rate data from the Ticker WebSocket feed, which provides relative_funding_rate and next_funding_rate_time for perpetual futures.

The interval field on FundingRateUpdate is None for Kraken because the ticker feed does not include a funding interval field and the Kraken API documentation does not specify a fixed funding period.

Rate limiting

The adapter implements automatic rate limiting to comply with Kraken's API requirements.

Reconciliation interval guidance

The execution engine's open_check_interval_secs and position_check_interval_secs settings create sustained REST API load that can exhaust Kraken's counter-based rate limit, especially on the Starter tier where the counter decays at only 0.33/sec. Each open-order check generates 1-3 REST calls (+1 or +2 counter each), and at short intervals the counter overflows before it can decay, causing EAPI:Rate limit exceeded errors.

Recommended settings for Kraken:

exec_engine=LiveExecEngineConfig(
    reconciliation=True,
    open_check_interval_secs=30.0,    # 30s minimum for Starter tier
    position_check_interval_secs=120.0,  # 2 minutes
)

Higher-tier accounts with faster counter decay can use shorter intervals. If you see EAPI:Rate limit exceeded errors in the logs, increase these intervals or reduce max_requests_per_second in the adapter config.

Configuration

The product types for each client must be specified in the configurations.

Data client configuration options

Execution client configuration options

For spot margin, default_leverage applies when an order has no per-order leverage param. margin_balance_asset only changes the TradeBalance summary denomination; per-position figures remain in the pair's quote currency.

Demo environment setup

To test with Kraken Futures demo (paper trading):

1. Sign up at https://demo-futures.kraken.com and generate API credentials.
2. Set environment variables with your demo credentials:
   • KRAKEN_FUTURES_DEMO_API_KEY
   • KRAKEN_FUTURES_DEMO_API_SECRET
3. Configure the adapter with environment=KrakenEnvironment.DEMO and product_types=(KrakenProductType.FUTURES,).

from nautilus_trader.adapters.kraken import KRAKEN
from nautilus_trader.adapters.kraken import KrakenEnvironment
from nautilus_trader.adapters.kraken import KrakenProductType

config = TradingNodeConfig(
    ...,  # Omitted
    data_clients={
        KRAKEN: {
            "environment": KrakenEnvironment.DEMO,
            "product_types": (KrakenProductType.FUTURES,),
        },
    },
    exec_clients={
        KRAKEN: {
            "environment": KrakenEnvironment.DEMO,
            "product_types": (KrakenProductType.FUTURES,),
        },
    },
)

Production configuration

The most common use case is to configure a live TradingNode to include Kraken data and execution clients. Add a KRAKEN section to your client configuration(s):

from nautilus_trader.adapters.kraken import KRAKEN
from nautilus_trader.adapters.kraken import KrakenEnvironment
from nautilus_trader.adapters.kraken import KrakenProductType
from nautilus_trader.live.node import TradingNode

config = TradingNodeConfig(
    ...,  # Omitted
    data_clients={
        KRAKEN: {
            "environment": KrakenEnvironment.LIVE,
            "product_types": (KrakenProductType.SPOT,),
        },
    },
    exec_clients={
        KRAKEN: {
            "environment": KrakenEnvironment.LIVE,
            "product_types": (KrakenProductType.SPOT,),
        },
    },
)

Dual-product configuration (Spot + Futures)

When trading both Spot and Futures markets, include both product types:

config = TradingNodeConfig(
    ...,  # Omitted
    data_clients={
        KRAKEN: {
            "environment": KrakenEnvironment.LIVE,
            "product_types": (KrakenProductType.SPOT, KrakenProductType.FUTURES),
        },
    },
    exec_clients={
        KRAKEN: {
            "environment": KrakenEnvironment.LIVE,
            "product_types": (KrakenProductType.SPOT, KrakenProductType.FUTURES),
        },
    },
)

Then, create a TradingNode and add the client factories:

from nautilus_trader.adapters.kraken import KRAKEN
from nautilus_trader.adapters.kraken import KrakenLiveDataClientFactory
from nautilus_trader.adapters.kraken import KrakenLiveExecClientFactory
from nautilus_trader.live.node import TradingNode

# Instantiate the live trading node with a configuration
node = TradingNode(config=config)

# Register the client factories with the node
node.add_data_client_factory(KRAKEN, KrakenLiveDataClientFactory)
node.add_exec_client_factory(KRAKEN, KrakenLiveExecClientFactory)

# Finally build the node
node.build()

API credentials

There are two options for supplying your credentials to the Kraken clients. Either pass the corresponding api_key and api_secret values to the configuration objects, or set the following environment variables:

When starting the trading node, you'll receive immediate confirmation of whether your credentials are valid and have trading permissions.

Contributing