Skip to content

Tutorial: a guarded swap with the SDK

This walks through a complete, runnable script that buys a token on testnet — with a spend cap you enforce yourself, a real slippage floor, and a balance check to confirm it worked. By the end you'll have the safe pattern you can build on.

You'll need: the SDK installed, a dedicated agent wallet funded on testnet, and a minted DID on that wallet.

The whole script

from decimal import Decimal
from primedelta import PrimeDelta, SwapSide, AccountStatus

RPC = "https://chain.testnet.primedelta.io"
SYMBOL = "AMMT1"
SPEND = Decimal("10")          # dUSD to spend
MAX_TRADE_USD = Decimal("25")  # your own per-trade cap
SLIPPAGE_BPS = 100             # 1%

pd = PrimeDelta(
    private_key="0x…",          # a dedicated agent wallet, funded on testnet
    web3_provider_url=RPC,
    network="testnet",
)

# 1. Authenticate.
pd.login()

# 2. Make sure this wallet can actually trade.
if pd.get_account_status() != AccountStatus.DID_MINTED:
    raise SystemExit("wallet has no minted DID — verify it first")

# 3. Enforce your own cap BEFORE quoting.
if SPEND > MAX_TRADE_USD:
    raise SystemExit(f"{SPEND} dUSD exceeds your cap of {MAX_TRADE_USD}")

# 4. Quote, and derive a real minimum-out (never zero).
quote = pd.quote_swap(SYMBOL, SwapSide.STABLECOIN_TO_STOCK, SPEND)
min_out = pd.min_out_from_quote(quote, slippage_bps=SLIPPAGE_BPS)
print(f"quote: {SPEND} dUSD -> ~{quote} {SYMBOL}, min accepted {min_out}")

# 5. Swap.
tx = pd.swap_exact_input(
    SYMBOL,
    SwapSide.STABLECOIN_TO_STOCK,
    amount_in=SPEND,
    min_amount_out=min_out,
)
print("swap tx:", tx)

# 6. Confirm it landed.
balances = pd.get_onchain_balances()
print(f"{SYMBOL} balance now: {balances.get(SYMBOL, Decimal(0))}")

What each step is doing

  1. Login signs a SIWE message with your signer and opens a session.
  2. The DID check is the gate: swaps require AccountStatus.DID_MINTED. Checking it yourself gives a clear error instead of a revert deep in the swap.
  3. The cap check comes before the quote. For a buy, the dUSD you spend is the trade's value, so you can reject an over-cap trade without pricing it — the same order the server uses.
  4. min_out_from_quote turns the quote into a floor: quote × (1 − slippage). Passing zero would accept any price; the swap rejects a non-positive floor anyway.
  5. swap_exact_input returns a transaction hash.
  6. get_onchain_balances returns a dict[str, Decimal]; the token appears once its balance is above zero.

Make it a sell

Flip the side and swap a token amount for dUSD:

quote = pd.quote_swap(SYMBOL, SwapSide.STOCK_TO_STABLECOIN, Decimal("2"))
min_out = pd.min_out_from_quote(quote, slippage_bps=SLIPPAGE_BPS)
tx = pd.swap_exact_input(SYMBOL, SwapSide.STOCK_TO_STABLECOIN,
                         amount_in=Decimal("2"), min_amount_out=min_out)

For a sell the value isn't known until the quote, so check your cap against the quoted dUSD out right after quoting.

Oracle stocks are different

A tokenized stock like AAPL has no AMM quote — check the market is open, read the oracle price, and compute the minimum-out yourself with extra slippage. See Place a swap.

Next → Watch a price and swap