https://sandbox-api.nuvante.ioStart here
Platform concepts
About a three-minute read. The role guides link here so they don't each have to explain the same ideas.
Non-custodial by design
Nuvanté coordinates transactions, and your assets stay with you the whole time. Every clearing participant and issuer brings their own wallet or custody setup, whether that's self-custody, MPC or a custody-as-a-service provider. We store a reference to it, such as an address, a registered handle or a delegation grant. We don't create wallets, hold private keys or act as a custodian.
Custody models explains how to run your custody. Chains explains how to set up your on-chain account.
The settlement saga
Every transaction runs as a coordinated sequence of steps, which we call a saga. The source asset is locked on-chain, the fiat leg is earmarked and settled through RTGS, the source token is burned, and for swaps the target token is minted. Either every leg completes, or the whole thing unwinds safely.
The public API only accepts two useCase values: redeem and swap. You tell us the outcome you want and the engine works out the steps.
Each use case has a walkthrough: Redemption and Stable-to-stable swap. For polling and terminal states, see Transaction lifecycle.
Derived settlement mode
You don't choose the order of steps in your request. The engine works it out from each issuer's declared capability profile, using fields like commitPoint, preBurnTransferRequired and feedbackChannel. That profile puts the issuer into one of three modes:
- Mode 1, delegated signing
- Mode 2, two-phase instruction
- Mode 3, single-phase with a pre-burn transfer
So two identical requests sent to two different issuers can take quite different routes and still end up in the same place.
The full field list is in Issuer capability declaration, and the modes are compared side by side in Settlement modes.
Proof-gated settlement
The engine only moves the RTGS leg forward once it has verified an on-chain burn proof, plus a mint proof for swaps. Issuer statuses like APPROVED or EXECUTING are just progress updates and don't mean anything has settled. Internally, we translate each issuer's own statuses into a common set of events. Instruction lifecycle lists them.
Reserve backing is checked against a declared policy (a backing ratio and a central-bank deposit ratio) using reserve proofs. A proof tells you whether the policy passed or failed and keeps the underlying figures private. Read Zero-knowledge reserve proofs for what this currently guarantees and what's still in progress.
Fees
Fees are added on top of the amount being sent. The receiver gets exactly the amount you state, and the fee is added to the sender's debit. We always show the fee as its own number. The initiate response carries three separate fields: receiveAmount, platformFeeAmount and debitAmount. Please keep them separate in your own UI and records too.
Platform fees covers the full rules and how the fee sweep works in RTGS.
