Order Types
Limit Orders
from kuru_sdk_py.manager.order import Order, OrderType, OrderSide
Order(
cloid="my-bid-1",
order_type=OrderType.LIMIT,
side=OrderSide.BUY,
price=100.0,
size=1.0,
)
Market Orders
await client.place_market_buy(quote_amount=100.0, min_amount_out=0.9)
await client.place_market_sell(size=1.0, min_amount_out=90.0)
Cancel Orders
Order(cloid="my-bid-1", order_type=OrderType.CANCEL)
Batch Updates
Combine limit orders and cancels into a single atomic transaction:
orders = [
Order(cloid="bid-1", order_type=OrderType.LIMIT, side=OrderSide.BUY, price=99.0, size=1.0),
Order(cloid="ask-1", order_type=OrderType.LIMIT, side=OrderSide.SELL, price=101.0, size=1.0),
Order(cloid="old-bid", order_type=OrderType.CANCEL),
]
txhash = await client.place_orders(orders, post_only=True)
Configuration
ConfigManager
The SDK uses ConfigManager to load configuration from environment variables with sensible defaults. See the Quick Start for the full list of environment variables.
import os
from kuru_sdk_py.client import KuruClient
from kuru_sdk_py.configs import ConfigManager
configs = ConfigManager.load_all_configs(
market_address=os.environ["MARKET_ADDRESS"],
fetch_from_chain=True,
)
client = await KuruClient.create(**configs)
ConfigPresets
Use presets for common scenarios:
from kuru_sdk_py.configs import ConfigPresets
preset = ConfigPresets.conservative() # Longer timeouts, more retries (production)
preset = ConfigPresets.aggressive() # Shorter timeouts, fewer retries (HFT)
preset = ConfigPresets.testnet() # Optimized for slower testnets
| Preset | Use Case |
|---|
conservative() | Production environments — longer timeouts, more retries |
aggressive() | High-frequency trading — shorter timeouts, fewer retries |
testnet() | Testnets — optimized for slower block times |
Production Guidance
Use a Dedicated RPC
The default public endpoints can be rate-limited. For production, use a dedicated RPC provider via RPC_URL and RPC_WS_URL.
Quote Cadence and Gas
Batch cancel/replace every second is expensive on-chain. Common approaches:
- Update only when mid price moves beyond a threshold
- Update at a slower cadence (e.g., 5-15s) unless volatility spikes
- Use fewer levels or smaller grids
- Enable EIP-2930 access list optimization (
KURU_USE_ACCESS_LIST=true)
Safety Checks
Always implement these safety checks before running in production.
- Stale data guard - Don’t quote if your market data feed is older than N milliseconds
- Min/max size - Ensure order sizes meet market constraints (or the tx will revert)
- Balance guard - Ensure margin balances can support your outstanding orders
- Circuit breakers - Stop quoting on repeated failures, disconnects, or extreme spreads. Use
await client.cancel_all_active_orders_for_market() to cancel all outstanding orders when the circuit breaker triggers
Running Examples
# End-to-end market making bot (requires PRIVATE_KEY + MARKET_ADDRESS)
PYTHONPATH=. uv run python examples/simple_market_making_bot.py
# Read-only orderbook stream (no wallet required)
PYTHONPATH=. uv run python examples/get_orderbook_ws.py
# Repeated batch placement + cancels
PYTHONPATH=. uv run python examples/place_many_orders.py
