How 1inch Exchange Finds Prices Across Decentralized Markets

The routing engine behind 1inch Exchange is called Pathfinder. Given a pair and a size, it searches for the combination of markets that returns the most output tokens, and it is allowed to split the order: part through one pool, part through another, the remainder through a two-hop path that passes via an intermediate token. Complex routes are common on large orders precisely because no single pool is deep enough to absorb them cheaply.

Splitting is not free. Every extra hop and every extra pool adds gas, so the cheapest route on paper can be the worse route once the transaction fee is counted. Pathfinder weighs expected output against estimated gas and picks the net-best result, which is why 1inch Exchange may return a plain one-pool route for a small trade and an elaborate multi-part route for a large trade in the very same pair minutes later.

A quote is an estimate, not a promise. Between the moment 1inch Exchange builds a route and the moment the transaction is included in a block, other trades can move the pools it planned to use. Slippage tolerance is the guard: it sets the minimum output the contract will accept, and if the market moves past it the swap reverts instead of filling at a bad price. Setting it very tight causes failed transactions that still cost gas; setting it very wide invites sandwich attacks. 1inch Exchange suggests a default, and an unusually wide setting should be treated as a warning rather than a convenience.

What the router weighs

• Available depth in each pool at the requested size
• Gas cost of each additional hop, which 1inch Exchange subtracts from the headline output
• Fees charged by the venues on the route
• Intermediate tokens that open a cheaper indirect path

Under the interface, 1inch Exchange is several protocols with different jobs. They share the same custody model but behave very differently at execution time.

Aggregation Protocol

This is the classic swap. The user pays the gas, the router contract executes the route in a single transaction, and the trade either fills or reverts within that block. It is the most direct mode on 1inch Exchange and the right choice when the priority is certainty about when the trade happens.

Limit Order Protocol

A limit order here is a signed message that only becomes executable when someone chooses to fill it at the stated price. Because the order lives off-chain until then, placing and cancelling it costs no gas and the filler pays for settlement. Orders on 1inch Exchange can carry extra conditions, which is how stop-style triggers and partial fills are built on the same primitive.

This is also the piece market makers integrate with directly through request-for-quote flows, letting professional liquidity reach 1inch Exchange users without ever sitting in a public pool.

Fusion

Fusion mode, introduced at the end of 2022, inverts the usual order of events. Rather than executing a route themselves, users sign an intent describing what they want swapped, and a network of resolvers competes to fill it through a Dutch auction in which the rate a resolver must beat improves over time. The winner submits the transaction and pays the gas, so a Fusion swap on 1inch Exchange can be gasless from the user's side, with that cost absorbed into the rate.

The tradeoff is time. A Fusion order fills when some resolver decides the auction has reached a profitable level, which typically takes seconds to a few minutes rather than one block, and in thin markets it can expire unfilled. In return, the order is not broadcast as an ordinary pending transaction, which narrows the surface for sandwich attacks. 1inch Exchange presents both modes side by side so the trader chooses the tradeoff deliberately.

Fusion+ and the wallet

Fusion+ extends the same idea across chains. A user signs one intent to move value from a token on one network into a token on another, and resolvers handle both legs, with escrow contracts and hashed secrets ensuring that either both sides settle or neither does. For the trader, a cross-chain swap through 1inch Exchange becomes a single signature instead of a bridge transaction followed by a second swap on the far side.

The 1inch Wallet is a separate self-custody mobile app with swapping built in. It stores keys on the device and connects to other applications, giving 1inch Exchange a first-party client, but it is not required: any standard wallet works with the web interface.

A swap carries several costs and only some of them belong to the aggregator. Network gas goes to the chain. Pool fees go to the liquidity providers on whichever venues the route touches. Price impact, the amount the trade itself moves the pools, is not a fee at all but behaves like one and grows with order size. What 1inch Exchange controls is the route, and a better route reduces the second and third of those.

On classic swaps, 1inch Exchange has historically not taken a protocol fee on top of those costs, though fee arrangements in decentralized products change through governance and product decisions, and third-party applications that embed 1inch Exchange through its API may add a fee of their own. In Fusion the economics differ by design: resolvers compete for the order and keep the difference between what a fill costs them and what they promised the user, so the cost lives inside the rate rather than appearing as a line item.

The practical habit is to read the quote screen rather than the marketing. Before signing, 1inch Exchange shows the expected output, the minimum guaranteed output at the current slippage setting, the estimated network fee, and the route itself. If the minimum sits far below the expected figure, the order is large relative to available liquidity, and breaking it into smaller trades or switching to a limit order is usually cheaper than forcing it through.

The number worth watching is not the headline rate but the minimum output. That is the only figure the contract actually guarantees.

Self-custody removes one category of risk and hands the trader another. No exchange can freeze or lose funds it never held, but key management, transaction review, and approval hygiene now belong entirely to the user. The contracts behind 1inch Exchange are audited, covered by a public bug bounty, and open to inspection, which is the standard defense in this category rather than a guarantee.

Most losses in this space happen around a protocol rather than through it. Phishing sites that imitate a familiar interface, careless token approvals, and fake support accounts have cost users far more than contract bugs. 1inch Exchange has also published security research affecting the wider ecosystem, including a 2022 disclosure about weak private keys produced by the Profanity vanity address tool.

Habits that prevent most incidents

• Bookmark the interface instead of reaching 1inch Exchange through a search advertisement
• Verify a token's contract address, since anyone can mint a token with a familiar name
• Review outstanding approvals periodically and revoke the ones no longer in use
• Treat any account offering 1inch Exchange support in a private message as fraudulent

1INCH is the network's governance token, distributed to early users in December 2020 and traded openly since. Holding it is not a requirement for trading: anyone can use 1inch Exchange without ever touching the token, and swaps never settle in it.

The token's role is voting and staking. Holders lock 1INCH to receive a staked position whose weight grows with the length of the lock, and that weight is used to vote in the 1inch DAO on protocol parameters, treasury spending, and resolver-related decisions in Fusion, which over time shape how 1inch Exchange routes and charges. Governance therefore sits with token holders rather than the founding team, though as in most DAOs participation concentrates among larger holders. For a trader, the sensible reading is that 1inch Exchange is a tool and the token is a separate decision entirely.

A large share of the activity never touches the website. 1inch Exchange exposes its routing through an API, and wallets, portfolio trackers, and other applications call it to offer swaps inside their own interfaces. The typical integration requests a quote, receives calldata for the router contract, and passes that to the user's wallet to sign, which keeps the custody model intact.

The appeal for builders is that liquidity aggregation is a tedious problem to rebuild. Rather than maintaining adapters for every new pool type, an application delegates the search to 1inch Exchange and inherits new venues as they are added. The protocols are also callable directly on-chain, so a contract can route through 1inch Exchange with no hosted service anywhere in the path.

1inch Exchange cannot conjure liquidity that does not exist. On a thin pair the best available route is still a poor price, and the aggregator's honest job is to display that plainly rather than disguise it. It also cannot reverse a signed transaction, cannot retrieve a token sent to the wrong address, and cannot move value between chains outside the Fusion+ flow.

Other omissions are deliberate. 1inch Exchange provides no custody, no margin, no first-party fiat on-ramp, and no account recovery, because each of those would require holding user funds or user identity. Traders who need them generally pair 1inch Exchange with separate services and accept that the exchange layer here is only the exchange layer.