If you still think cross-chain intent protocols are just a faster kind of bridge, you are going to pick the wrong one. Bridges lock an asset on chain A and mint a representation on chain B. Intents skip the asset plumbing entirely: a user signs a desired outcome, a solver network competes to deliver it, and settlement happens only if the outcome is met. Same problem on the surface, completely different machine underneath. This 2026 ranking compares the ten intent protocols that matter, scored on ERC-7683 conformance, solver economics, and whether the protocol is stablecoin-native or DEX-swap-native. You will come away knowing which one wins for your use case, not a vague "overall best."

Bridges and intents solve the same user problem (value on chain A ends up on chain B), but they use opposite settlement models. A bridge is a custody contract plus a message layer: you deposit into an escrow, a relayer sees the event, a minter releases a wrapped token on the destination. The bridge has to hold liquidity, manage risk on both sides, and carry the exploit surface of its message layer. An intent is closer to a limit order than a transfer. The user signs "I will pay 1,000 USDC on Base and I want 1,000 USDC on Arbitrum within 30 seconds." A competitive set of solvers sees the intent, one of them fronts the output from its own inventory, and it claims the user's input after proving delivery.

The implication matters for anyone integrating cross-chain flow. With a bridge, the user waits for the bridge's finality and trusts the bridge's security model. With intents, the user gets a signed commitment that either completes in full or never executes at all. Paradigm's original framing of intents is still the clearest articulation: bridges move bytes, intents match outcomes. The intent-based architecture treats the solver market as the execution layer, not the bridge.

Every protocol below is scored on three axes that actually determine which one you should use.

1. ERC-7683 conformance. The ERC-7683 cross-chain intents standard defines a common CrossChainOrder struct and resolver interface so solvers can fill intents from any compliant origin. Conformant protocols get access to a wider solver pool and can interoperate; non-conformant protocols keep their solver network proprietary. Conformance is not binary — some protocols support 7683 at the order-emission layer but keep proprietary settlement, others are native.

2. Solver economics. The margin a solver earns determines how many of them will show up, and how many show up determines how tight your spreads are. Stablecoin 1:1 flows have razor-thin margins but high frequency. DEX swaps carry wider spreads but more execution risk. Protocols that do not give solvers a healthy spread end up with one dominant solver, which is a bridge with extra steps.

3. Stablecoin-native vs DEX-swap-native. A protocol built to move USDC between L2s optimizes for speed, atomicity, and amount fidelity. A protocol built to swap ETH for a long-tail token optimizes for price discovery and MEV resistance. These are different products. Ranking them against each other on one dimension is how you end up recommending UniswapX for treasury payouts — which would be a mistake.

Eco is a stablecoin execution network that uses intents as its core primitive. It is not "an intents protocol" in the Anoma sense — intents are how it moves stables, not the product itself. That distinction shows up in the design: Routes is built for 1:1 stablecoin movement across 15 chains (Ethereum, Optimism, Base, Arbitrum, HyperEVM, Plasma, Polygon, Ronin, Unichain, Ink, Celo, Solana, Sonic, BSC, Worldchain) with seven stablecoin pairs (USDC, USDT, USDC.e, oUSDT, USDT0, USDbC, USDG). Users sign a desired outcome, solvers compete to fulfill it, and execution is atomic.

Two live route configurations make this concrete. The Native Route uses CCTP under the hood for chains that support it, giving native-USDC finality without wrapped assets. The Hyperlane Route uses Hyperlane for message passing on chains where CCTP is not yet available. Both feed the same solver network, so integrators write one call and the protocol picks the right rail. Developers can publish a cross-chain intent with a single API call.

Wins for: stablecoin treasury operations, payment rails, agentic payouts, anything where the user wants dollars-in / dollars-out at 1:1. Not for: long-tail token swaps or DEX-style price discovery.

UniswapX is the benchmark for DEX-intent protocols. Users sign a Dutch-auction order that starts at the best price a filler is willing to offer and decays to the user's floor. Fillers (Uniswap's term for solvers) compete off-chain and settle onchain. The architecture is optimized for price discovery on volatile pairs and MEV resistance — the Dutch auction is a clever way to let the winning filler capture MEV rather than having a searcher sandwich the user.

The UniswapX model works beautifully for ETH/USDC, ETH/WBTC, and similar liquid pairs. It is less suited to stablecoin-to-stablecoin at 1:1 — you end up paying auction-decay spread for a flow that does not need price discovery. Wins for: single-chain and cross-chain DEX swaps on liquid pairs. Not for: payments or 1:1 stablecoin moves.

CoW Protocol pioneered batch auctions for DEX intents: instead of filling each order individually, CoW batches them, finds coincidences of wants (two users trading opposite directions clear against each other at a uniform clearing price), and only routes the residual to external liquidity. The solver competition is structured around who can return the most surplus to users in a single batch.

CoW is still primarily Ethereum-mainnet and EVM L2 focused for spot swaps. Its cross-chain story is building but is not the product's center of gravity in 2026. Wins for: MEV-protected DEX swaps with coincidence-of-wants savings. Not for: payments, stablecoin rails, or chains outside EVM.

Across is the closest thing to a pure bridge competitor on this list. It uses an intent-style model (users deposit on the origin, relayers front the asset on destination, settlement happens later via a dispute window), and it conforms to ERC-7683. Across's strength is speed on major EVM L2s for ETH, WETH, and the big stablecoins, with a mature relayer network and a strong security track record.

The trade-off versus a pure intent protocol like Eco is that Across still lives partly in bridge-land — its canonical settlement is the UMA optimistic oracle, not a solver-market clearing. For stablecoin-specific flows, a stablecoin-native intent stack will usually have tighter economics. Wins for: fast EVM-to-EVM asset transfers where you want the bridge-speed without lock-and-mint risk. Not for: non-EVM destinations or stablecoin-first treasuries.

Aori takes a market-making-first approach. It is an orderbook for cross-chain intents where professional market makers quote two-sided liquidity and users match against those quotes. Aori is explicit that it is building for market makers first and users second — the thesis is that if you give MMs the right primitives (canonical intent format, atomic settlement, capital efficiency), user flows follow.

In practice this means Aori has some of the best execution for institutional-size cross-chain trades where an RFQ-style quote beats an auction. The Aori model is not trying to be the consumer default. Wins for: MM-driven flows, RFQ on liquid pairs, professional liquidity sourcing. Not for: simple retail swaps or small stablecoin transfers.

1inch Fusion retrofitted intent semantics onto 1inch's aggregator. Users sign a gasless order, resolvers (1inch's term for solvers) compete via Dutch auction, and the user pays no gas while receiving the best aggregated price across 1inch's DEX graph. Fusion is strong specifically because it sits on top of 1inch's existing liquidity graph across 10+ chains.

It is less a standalone cross-chain intent protocol and more an intent-flavored aggregator. Cross-chain Fusion is expanding but the protocol's primary value is still single-chain gasless swaps with best-price routing. Wins for: gasless single-chain DEX swaps with aggregator-grade routing. Not for: true cross-chain stablecoin payments.

dYdX's v4 chain and emerging intent layer are optimized for perpetual futures. Users can express intents around leverage, liquidation thresholds, and funding — things generic cross-chain intent protocols do not model. For traders routing collateral in and executing perps at the same time, dYdX's vertically integrated stack is competitive. For anything that is not perps, it is the wrong tool.

Wins for: perp trading with collateral-routing intents. Not for: anything outside perps.

Anoma is the maximalist take. Rather than building an intent protocol for a specific use case, Anoma is building a general-purpose intent machine — a chain where all transactions are intents, composition is native, and the solver market is the protocol. Anoma's design is the most conceptually interesting on this list, but as of 2026 it is still an emerging ecosystem. Production users are researchers and intent-native application builders rather than mainstream integrators.

Wins for: research, novel intent-native applications, teams comfortable on new infrastructure. Not for: production stablecoin flows or conservative integrators today.

Symm focuses on symmetrical derivatives and intent-style liquidity for structured products. It is a niche but technically strong implementation — if your product involves symmetrical payoffs or structured derivatives, Symm is one of the few intent stacks that models them natively. For general cross-chain payments, Symm is out of scope.

Wins for: symmetrical derivatives, structured products with intent settlement. Not for: payments, swaps, or general cross-chain transfers.

Everclear (formerly Connext) is not a user-facing intent protocol — it is a netting layer underneath them. Everclear coordinates solver inventory across chains so that solvers do not need to hold idle liquidity on every destination. For integrators, it is relevant because the solver you are relying on downstream is probably rebalancing through Everclear. For end-user intent flow, you rarely interact with it directly.

Wins for: solver-infrastructure teams, protocols building their own solver networks. Not for: direct user-facing intent integration.

The single most common mistake integrators make is picking a DEX-swap-native intent protocol for a stablecoin payment rail, or vice versa. The two categories optimize for different things and the resulting user experience is different.

Stablecoin-native protocols (Eco Routes, Across for stables, Circle CCTP as the underlying rail) optimize for amount fidelity — you send 1,000 USDC on origin, you receive as close to 1,000 USDC on destination as solver economics allow, minus a fixed protocol fee. The user should not see price slippage because there is no price. Speed and atomicity are the metrics.

DEX-swap-native protocols (UniswapX, CoW, 1inch Fusion) optimize for price discovery — the user is selling asset A for asset B at a market-clearing price. Solver competition compresses the spread but there is always spread. Slippage protection and MEV resistance are the metrics.

Alt text for table: "Cross-chain intent protocols 2026 comparison by category, ERC-7683 conformance, and best use case."

A solver's job is to front the output of an intent from its own inventory and get paid a small spread when it settles. The spread size, the settlement speed, and the risk (price, fraud, bridge) determine who shows up. Healthy solver markets have many competitors, tight spreads, and short settlement windows. Unhealthy markets have one dominant solver and nothing you would call a market.

For stablecoin 1:1 flows, the spread is a few basis points, the risk is near-zero because the asset is the same on both sides, and settlement is typically measured in seconds. That is why stablecoin-native protocols tend to have dense solver participation — the risk/reward is legible. For DEX swaps, the spread is wider but so is the execution risk (price moves during the auction). This is why solver concentration is higher in DEX-intent protocols — the capital and sophistication required to price volatile assets in real time is not commodity.

If you are choosing a protocol for production use, ask the team how many active solvers they have, what the median spread is on your specific asset pair, and what happens when their top solver goes offline. Two of those questions will tell you more than any whitepaper.

The ERC-7683 standard defines a canonical CrossChainOrder format and a resolver interface, so any solver implementing 7683 can fill orders emitted from any 7683-compliant origin. In practical terms this means a solver that already fills Across orders can also fill Eco-emitted orders with minimal additional work, which deepens the liquidity available to every protocol that implements the standard. Proprietary solver networks become strictly worse — why run a private network when the public one already has your competitors' solvers plus yours?

In 2026 the question is no longer "does your protocol support intents" but "does your protocol's order format speak ERC-7683." If a protocol's answer is no, its solver network is capped at whoever the team has recruited directly. If the answer is yes, the addressable solver set is the entire 7683 universe. That is a category-redefining difference.

Stop asking which intent protocol is "best." Start asking which one wins for your specific job.

If you are integrating cross-chain stablecoin flow into a product and you want one call that fans out to the right rail per chain, Eco's Routes CLI is the shortest path. Teams deploying stablecoin payments in production often combine the two live route configurations behind a single integration: Native Route for CCTP-supported pairs, Hyperlane Route for the rest, same solver set, same API.

A bridge locks an asset on the origin chain and mints a representation on the destination. An intent protocol lets users sign a desired outcome and lets solvers front the output from their own inventory, settling later. Bridges manage custody and message-passing risk; intent protocols rely on solver competition and atomic settlement. See the full intent-based architecture explainer for the mechanics.

As of 2026, Eco Routes, Across, and Aori are ERC-7683 compliant. UniswapX, CoW, and 1inch support partial conformance at the order-emission layer. Anoma is a native superset. Non-conformant protocols cap their solver pool to whoever the team has recruited directly.

Eco is a stablecoin execution network that uses intents as a primitive, not an "intents protocol" in the general-purpose sense. The product is 1:1 stablecoin movement across 15 chains with atomic settlement, with intents as the mechanism. The solver network clears flows rather than a bridge contract.

Solvers earn a spread between the user's input amount and the output they front. Protocols typically charge a protocol fee on top, or take a share of solver revenue. Healthy intent markets have many solvers competing the spread down, so protocol economics depend on volume rather than margin.

Generally yes, because intents are atomic — the intent either completes in full or never executes at all. Bridges have a larger exploit surface (message layer, escrow, wrapped-token minter) and locked-liquidity exposure. That said, solver solvency and oracle risk still apply, so always check a protocol's specific security model rather than assuming "intents means safe."

Yes — stablecoin-native intent protocols are production-grade in 2026. Teams moving seven- and eight-figure stablecoin volumes daily are live on Eco Routes, Across for stables, and Circle's CCTP as an underlying rail. For any of these, you can publish a cross-chain intent with a single integration call.