Testing on testnet

Test your Aquarius integration on Stellar testnet before touching real funds.

Aquarius runs a full deployment on Stellar testnet — the AMM router, pools, and a dedicated instance of the backend API. Everything in these docs works identically on testnet; only the constants change.

Setup

• Constants — the testnet router, API base URL, and RPC/Horizon endpoints are listed in Addresses & networks. The Quickstart is pre-configured for testnet.

• Funding accounts — testnet XLM is free from friendbot: https://friendbot.stellar.org?addr=<PUBLIC_KEY> funds any new account with 10,000 XLM.

• Test assets — testnet versions of AQUA, USDC, USDT and others are issued by the shared testnet issuer GAHPYWLK6YRN7CVYZOO4H3VDRZ7PVF5UJGLZCSPAEIKJE2XSWF5LAGER. Add a trustline, then acquire them by swapping testnet XLM through the AMM.

What to expect

• Thin liquidity. Testnet pools hold test liquidity — quotes and slippage won't match mainnet. Use testnet to verify your integration logic, not your pricing assumptions.

• You can create pools. Pool creation is permissionless on testnet too, so you can set up exactly the pool configuration your tests need.

Testnet resets

Stellar's testnet is reset by SDF 2–4 times per year (17:00 UTC, announced at least two weeks in advance) — the ledger is restored to genesis: all accounts, trustlines, claimable balances, and contract state are wiped. The network passphrase stays the same.

After each reset, Aquarius rebuilds its testnet stack. What survives and what doesn't:

Survives a reset	Wiped by a reset
Contract addresses (router and core contracts) — redeployed deterministically: pinned deployer seed + fixed salts + the unchanged network passphrase	All accounts, balances, and trustlines — including your test accounts
Test-asset identities (AQUA:GAHP…) — the issuer keys are kept and re-provisioned	Claimable balances — votes and locks
Endpoints — Horizon, Soroban RPC, the testnet API URL, the network passphrase	Pools and their liquidity — recreated during the rebuild
	Backend-indexed data — re-indexed after the rebuild

The rebuild takes time — right after a reset, expect failures until the stack is back, even though the addresses you're calling are unchanged.

If your integration suddenly fails against testnet:

1. Check whether a reset just happened (announced in advance by SDF).

2. Re-fund your test accounts via friendbot and re-create trustlines.

3. If it still fails, the rebuild may not be finished — verify against Addresses & networks or ask in Discord.