Intent execution.
Outcomes, not transactions.
Intent execution lets an application declare what a user wants to achieve and hand the entire transaction pipeline to Eraivo — gas, routing, bridging, signing, and settlement included.
A transaction tells a chain exactly what to do. An intent declares the desired outcome — "swap 1.5 ETH for SOL", "move USDC to the cheapest venue" — and leaves the execution path open.
Eraivo resolves that intent into a concrete, multi-leg execution plan, verifies it off-chain, and only then dispatches it. The developer never hand-builds chain-specific calldata.
Every intent moves through the same four stages before it touches a chain:
- —Declare — the SDK accepts a high-level outcome, not calldata.
- —Verify — safety simulation, slippage bounds, and risk screening run off-chain.
- —Execute — non-custodial relayers orchestrate every leg atomically.
- —Prove — on-chain proofs are indexed and surfaced in real time.
How is an intent different from a transaction?
A transaction specifies exact on-chain operations. An intent specifies the outcome and lets Eraivo choose the route, chain, and sequencing.
What happens if part of an intent fails?
Multi-leg execution is atomic. If any leg fails, the whole operation rolls back — there are no partial states or stuck funds.
Does the user give up custody?
No. Signing is non-custodial and KMS/HSM-backed; key material never leaves the enclave.