Cross-Chain Intents: How ERC-7683 Replaces Bridges
Intent-based transfers let you say what you want and pay a solver to deliver it. Here is how ERC-7683 works and where the risk actually sits.
For years, moving assets between chains meant locking tokens in a bridge contract and praying the bridge did not get hacked. Intent-based transfers flip that model: you sign what you want, and a competitive market of solvers races to deliver it. ERC-7683 is the standard trying to make that the default.
Quick answer
ERC-7683 is a shared Ethereum standard for cross-chain intents, co-authored by Uniswap Labs and Across. Instead of manually bridging, you sign an order that specifies what you are paying on the source chain and what you want on the destination chain. A solver fronts the destination-chain funds instantly, then settles the source-chain payment for itself. It is faster and usually cheaper than a classic lock-and-mint bridge, but your funds still depend on the underlying verification layer the settlement contract trusts.
Key takeaways
- ERC-7683 standardizes cross-chain intents, so many apps can share the same solver network.
- You express a desired outcome; a solver handles the multi-step execution and takes a fee.
- Delivery is often near-instant because the solver fronts destination funds before final settlement.
- The standard does not invent its own security; it relies on a bridge or verification layer underneath.
- Observed fees on mature routes ran roughly 0.05 to 0.15 percent, cheaper than many pooled-liquidity bridges.
What an "intent" actually is
A traditional transaction spells out every step: approve this token, call that contract, wait for this confirmation. An intent instead describes the goal. You sign a message that says, in effect, "I will give up 1,000 USDC on Arbitrum, and I want at least 999 USDC on Base within two minutes." You do not specify how it happens.
That signed order contains a few concrete fields:
- Source-chain inputs: what you are paying and on which chain.
- Destination-chain outputs: the minimum you will accept and where.
- A settler contract address: the contract that finalizes the deal.
- A fill deadline: after which the order expires.
- An order data hash: a commitment binding the terms together.
A solver (sometimes called a filler or relayer) sees your order off-chain, simulates whether filling it is profitable, and if so delivers the output on the destination chain almost immediately. It then claims your source-chain input as repayment plus a small spread.
Bridges versus intents
The practical difference is who takes on the waiting and the risk. A classic bridge makes you wait for cross-chain messages before you receive anything. An intent solver eats that delay for you and gets paid to do it.
| Aspect | Classic lock-and-mint bridge | ERC-7683 intent |
|---|---|---|
| What you do | Execute a multi-step bridge flow | Sign one order describing the outcome |
| Speed | Wait for cross-chain finality | Solver fronts funds, often seconds |
| Who waits for finality | You | The solver |
| Fee model | Bridge fee, sometimes fixed | Competitive solver spread |
| Failure mode | Bridge exploit, stuck funds | Order unfilled, then refunded |
| Standardization | Bridge-specific | Shared across compatible apps |
The shared-standard part matters more than it sounds. Because ERC-7683 defines a common order format, a solver built for one app can fill orders from another. As of early 2026, orders were being settled by fillers including Across, Eco, LI.FI, UniswapX, and Symbiosis, and the Ethereum Foundation's Open Intents Framework pulled dozens of teams toward the same plumbing.
Where the security really lives
Here is the part marketing pages gloss over: ERC-7683 does not solve cross-chain verification. It assumes a verification primitive already exists between the two chains and leaves the choice to whoever deploys the settler contract. Your funds ultimately depend on that layer, not on the intent standard itself.
Two broad settlement styles are common:
- Optimistic settlement. The solver claims it filled your order, a challenge window opens, and if nobody disputes it within the window the solver is paid. This is how Across-style systems work. It is capital-efficient but introduces a delay before the solver is made whole, and it leans on honest challengers.
- Verification-backed settlement. The settler relies on a messaging or verification layer such as a light-client bridge or an interoperability protocol to confirm the fill before releasing funds. Here your risk collapses into that layer's security.
So the honest question to ask any intent product is not "is ERC-7683 safe" but "what verifies the fill, and what happens if that verifier is wrong or compromised." That is the same underlying concern behind our explainer on why crypto bridges keep getting hacked, because the settlement layer under an intent is often a bridge wearing a nicer interface.
What can go wrong for you
Most bad outcomes are annoyances rather than catastrophes, but they are worth knowing before you sign.
| Risk | What happens | How it is handled |
|---|---|---|
| No solver fills your order | Nobody finds it profitable | Order expires and you are refunded the input |
| Slippage on output | You receive the minimum you set | You control this by setting the minimum output |
| Settler or verifier exploit | The layer confirming fills is broken | This is the real tail risk; depends on the underlying design |
| Signing a malicious order | A phishing site crafts a bad intent | Read the typed data; reject unfamiliar settlers |
The last row is the one that overlaps with everyday self-custody hygiene. An intent is still a signature, and a fake site can present an order that pays a solver you never intended to enrich. The defenses are the same ones we cover in wallet drainer scams and token approvals: reach apps from your own bookmarks and never approve a signature you cannot read in plain language.
What to do right now
If you are about to use an intent-based transfer:
- Confirm the app is legitimate. Navigate from your own bookmark, not a search ad or a DM.
- Check the minimum output. Set a slippage tolerance you can live with; the solver only guarantees the minimum you signed.
- Ask what settles the fill. Prefer products that clearly document their verification or optimistic-challenge design.
- Start small. Move a test amount before routing a large transfer, exactly as you would with any new route.
- Read the signature. Reject any order with an unfamiliar settler contract or fields you cannot interpret.
For a broader view of how transactions finalize under the hood, our guide to blockchain finality and reorgs explains why the source-chain confirmation a solver waits for is not always as final as it looks.
Frequently asked questions
Is an intent transfer safer than a bridge?
It removes some user-facing steps and shifts the finality wait onto a solver, but it does not remove bridge risk. The settlement contract still relies on some verification layer, and that layer is where the serious risk lives. Treat "intent" as a better interface, not automatic safety.
Why is it usually cheaper?
Solvers compete to fill your order and pool liquidity across many routes, so spreads compress. On mature stablecoin routes, observed fees sat around 0.05 to 0.15 percent, below what many bridges that do not share solver liquidity charged.
What happens if no solver fills my order?
Nothing is lost. If your order is unprofitable or expires past its fill deadline, you are refunded the input on the source chain. The main cost is the time you waited.
Do I still need to approve tokens?
Usually yes, at least once, so the settler contract can pull your input. Approve the exact amount when you can rather than an unlimited allowance, to limit exposure if the contract is ever compromised.
This article is for general information and is not financial advice.
