Swaps on Avalanche - SODAX Docs

SODAX swaps are intents, not bridges. Your user signs one transaction on Avalanche that locks the input and declares what they want on the destination network. Solvers fill it; the SODAX relayer carries proof between Avalanche, the hub (Sonic), and the destination. If nothing fills, the intent can be cancelled and funds recovered. From your side it is one SDK call. From your user’s side it is one wallet prompt, plus an approval the first time they swap an ERC-20 token.

What this looks like on Avalanche

Source: the intent transaction is an ordinary EVM transaction on the C-Chain, signed by the user’s wallet. ChainKeys.AVALANCHE_MAINNET ('0xa86a.avax') identifies the network everywhere in the SDK.
Reach: any solver-compatible asset on any SODAX network is a valid destination, 166 assets today. The on-Avalanche asset list only bounds what users hold on Avalanche itself, not what they can swap into.
Tokens: AVAX uses the zero address; every other token is a standard C-Chain ERC-20 contract. Read addresses from the SDK config rather than copying them.
Settlement: output lands on the destination network at dstAddress. Swaps into Avalanche settle as ordinary AVAX or ERC-20 balances at the recipient’s C-Chain address, with no extra step on the receiving side.

Approvals

swap() does not approve the input token for you. AVAX is the native token and needs no approval. Every ERC-20 does: SODAX’s asset manager pulls it through the token’s ERC-20 interface, so check the allowance first and approve when it falls short.

const allowance = await sodax.swaps.isAllowanceValid({ params, walletProvider });
if (!allowance.ok) throw allowance.error;

if (!allowance.value) {
  const approval = await sodax.swaps.approve({ params, walletProvider });
  if (!approval.ok) throw approval.error;
  // wait for the approval to be mined before calling swap()
}

params is the same intent params object you pass to swap().

Quoting

The solver API quotes both directions:

const quoteResult = await sodax.swaps.getQuote({
  token_src: avax.address,            // token on Avalanche
  token_dst: usdc.address,            // token on the destination network
  token_src_blockchain_id: ChainKeys.AVALANCHE_MAINNET,
  token_dst_blockchain_id: ChainKeys.ARBITRUM_MAINNET,
  amount: 1n * 10n ** BigInt(avax.decimals), // 1 AVAX
  quote_type: 'exact_input',          // the only supported quote type
});

getQuote deducts any configured partner fee before quoting, so quoted_amount is the net output the user actually receives.

Fees

Using SODAX is free to integrate. Two fees can apply to a trade:

Base fee: a fixed 0.1% of the input, taken by the protocol. Not configurable; compute it ahead of time with getSolverFee(inputAmount).
Your fee: optional platform fee on top, set by you and paid to you. It is the only fee you control. Configure it at SDK setup and check it with getPartnerFee(inputAmount). See Monetize SDK.

The user also pays the C-Chain network fee for the intent transaction, in AVAX, like any other C-Chain transaction.

Executing

Prefer swap(). It creates the intent on Avalanche, verifies it landed, submits it to the relay, waits for the packet on the hub, and notifies the solver. The Quickstart has the complete call. Track progress with getStatus(request), and cancel unfilled intents with cancelIntent. Limit orders (intents without deadlines) work from Avalanche too, via createLimitOrder. Orchestrating manually with createIntent and submitIntent works the same as on any EVM network: Avalanche intents need no relay extra data.

Error handling

Operations return Result<T> instead of throwing. The core methods (swap, createIntent, postExecution) return a typed SodaxError union you can switch on by error.code.

Full reference, including raw mode, limit orders, cancellation, and the complete error-code table: Swaps.

​What this looks like on Avalanche

​Approvals

​Quoting

​Fees

​Executing

​Error handling