Cross-chain swap slippage sets a floor on output

Cross-chain swap slippage tolerance sets the minimum output a route may deliver before execution is rejected or handled under its refund rules. Across and LI.FI documentation describe tolerance through a minimum output amount, while the route may include swaps on one or both chains as well as the transfer between them.

The setting is useful only when you know which parts of the route it protects and what happens if a later leg misses its minimum.

A quote estimates output from current prices and route fees; the tolerance defines how far execution may move from that estimate. For an exact-input trade, a 1% tolerance generally means the output floor is 99% of the quoted output, after the quote’s stated costs. It does not guarantee that execution will occur at the estimate. It sets a limit on what the route can accept.

Route selection is a separate decision from that floor: the guide to manta bridge covers how to compare paths across chains. A route may bridge the same token, or swap before or after the transfer; those choices affect fees, execution conditions and which swap legs need a slippage bound.

What does slippage tolerance control on a cross-chain route?

It controls the permitted difference between a quoted swap output and the output accepted by the swap leg. LI.FI’s quote documentation exposes an estimated toAmount and a toAmountMin, the minimum value the transfer is expected to yield including slippage. That minimum is the useful figure to inspect: a percentage alone does not show the token amount at risk.

On a route with an origin-chain swap and a destination-chain swap, there may be more than one floor. Across documents an example in which a 1% tolerance is divided equally between the two legs, giving each a 0.5% bound. That is one provider’s implementation, not a universal rule. A route interface can apply a global tolerance differently, or expose leg-specific settings.

Why can the destination output differ from the quote?

Each swap executes against liquidity on its own chain, and pool prices can change between quoting and execution. A destination swap also occurs after the cross-chain transfer, so its execution is exposed to conditions at that later point. Bridge fees and transfer mechanics affect the final amount too, but they are not themselves slippage; compare the quoted net output and its minimum rather than treating every deduction as price movement.

A tight floor limits how much worse a swap may execute, but it can cause that leg to fail when the market moves beyond the bound. A wider floor allows more price movement and makes execution more permissive. On a cross-chain route, failure handling depends on the route’s design: a source transaction may already have completed, and the destination leg may have a refund path rather than reverting the source transaction. Check the quoted minimum and the route’s stated handling of failed destination swaps before signing.

How should you set and check the tolerance?

Start with the actual minimum output shown for the route, then check which legs it covers and whether the interface offers automatic or manual tolerance. Across’s API, for example, accepts an automatic setting or a numerical percentage; its documentation describes how a numerical setting is split across legs. A displayed percentage without that context is incomplete.

• Compare the quoted output with the minimum output in the destination token.
• Check whether the route swaps before the bridge, after it, or on both sides.
• Confirm whether the tolerance applies to each leg or is allocated across the route.
• Read how the route handles a failed destination swap, including where any refund is sent.

For most users, the better choice is the narrowest tolerance that still gives the quoted route room to execute under its expected conditions. Widen it only when the minimum output remains acceptable and the route’s failure handling is clear; a larger tolerance changes the price floor, not the route’s fees or bridge settlement.