Chain Abstraction Explained: Intents and Solvers

Chain abstraction is the idea that you shouldn't need to know, or care, which blockchain your assets are on. You don't bridge, switch networks and buy gas tokens by hand. You state the outcome you want, and infrastructure handles the routing. Today it mostly runs on intents and solvers, plus smart accounts that unify balances.

The term gets stretched to cover almost anything multichain, so this guide sticks to the mechanics: what the main building blocks are, how intents differ from traditional bridges, what you're trusting when a solver fills your order, and where the approach still falls short.

What chain abstraction means

NEAR's developer documentation defines chain abstraction as a set of technologies that let applications work across many blockchains while hiding each chain's complexity from both developers and end users. In practice, that complexity shows up as friction most crypto users know well:

• Separate balances on each network, even for the same asset.
• A different gas token on every chain, so you can't move USDC on a new network until you've bought some of its native coin.
• Multi-step bridging, with approvals, waiting periods and wrapped tokens that may not be the version a given app accepts.
• Different address formats and wallets for Bitcoin, EVM chains, Solana and others.

Chain abstraction tries to collapse those steps into one action. It isn't a single protocol. Several independent components each hide one part of the problem.

The building blocks

Component	What it hides	Examples
Intents and solvers	Route selection, liquidity sourcing, execution steps	NEAR Intents, Across, UniswapX
Unified smart accounts	Per-chain balances and gas tokens	Particle Network Universal Accounts
Multichain signing	Separate keys and wallets per chain	NEAR Chain Signatures
Shared standards	Custom integrations for every solver and protocol	ERC-7683 (draft)

Intents are signed statements of what you want, not how to get it, for example "I have 1 ETH on Ethereum and want USDC on Arbitrum at this address." Solvers, also called market makers, fillers or relayers depending on the protocol, compete to deliver that result.

Unified accounts attack the balance problem. Particle Network describes its Universal Accounts as smart-contract accounts built on ERC-4337 that present one account and one balance across chains. Transactions draw on your combined balance, and assets are bridged automatically when needed. Its docs also describe paying gas with whichever supported token you hold, and an EIP-7702 mode that lets an existing externally owned account use the system without moving funds.

Multichain signing lets one account control addresses elsewhere. NEAR's Chain Signatures allow NEAR accounts and smart contracts to sign transactions for other chains, including Bitcoin, Ethereum and Solana.

Intents vs bridges

A classic lock-and-mint bridge locks your asset in a contract or multisig on the source chain, and a representation is minted on the destination. That pool of locked collateral is the weak point. In August 2022, Chainalysis estimated that about $2 billion had been stolen across 13 cross-chain bridge hacks that year, 69% of all funds stolen in 2022 up to then. The Ronin bridge alone lost roughly $625 million in March 2022 after an attacker gained control of five of its nine validators.

Intent-based systems change who carries the risk. Instead of minting a claim, a solver delivers the real asset on the destination chain from its own inventory, then gets repaid from your deposit on the source chain later.

Across is a clear example. Its docs describe relayers who fill your deposit on the destination chain with their own tokens. Off-chain dataworkers bundle those fills and propose them at least every 1.5 hours or so. Each bundle goes through UMA's optimistic oracle: the proposer posts a bond, anyone can dispute during a challenge window, and disputes go to a vote of UMA token holders. Across says the system needs only a single honest party willing to dispute an invalid proposal.

NEAR Intents settles differently. Your intent goes out as a quote request, market makers bid, and an approved quote is executed atomically by the Verifier contract on NEAR (intents.near). Either the swap completes or you're refunded. Balances inside the contract are tracked in an on-chain ledger, and withdrawals to external chains go through token bridges.

	Lock-and-mint bridge	Intent-based swap
What you receive	A wrapped or minted representation	Usually the destination chain's real asset
Who fronts liquidity	The bridge's locked collateral	A solver's own inventory
Main risk	Compromise of the locked pool or its signers	Settlement contract, oracle or bridge bugs; solver no-shows
Speed driver	Source-chain finality plus message relay	How quickly a solver is willing to fill
Pricing	Protocol fee plus gas	Competitive quotes from solvers

Solvers: what you're actually trusting

Solvers are professional market makers. They hold inventory across chains, quote prices, and absorb finality risk. If the source chain reorganizes after they've paid you, that's their problem, which is why they price it into the quote. For the user, the trust picture usually looks like this:

• The settlement layer has to be correct. Whether it's NEAR's Verifier contract or Across's spoke pools and hub pool, a bug there affects everyone using it.
• Dispute or oracle systems have to work. Optimistic designs assume someone is watching and willing to challenge a bad proposal.
• Bridges haven't disappeared. In NEAR Intents, withdrawals to external chains still rely on token bridges. In Particle's model, assets are bridged automatically behind the scenes.
• Someone has to show up. If no solver takes your order, you wait for a refund or a slower fallback. Across, for example, can use liquidity provider capital for "slow fills" when no relayer acts.

Solver infrastructure can also fail on its own. In October 2025, the Bitcoin swap protocol Garden Finance lost about $11 million after one of its solvers was compromised. Garden said user funds weren't affected, but it took its app offline. In July 2026, roughly $450,000 was drained from Garden contracts after an attacker breached an independent solver's off-chain database, and Garden again said only solver-owned assets were hit. Both cases show the design working as intended for users while the service still went down.

ERC-7683: a common language for solvers

One structural weakness of intents is fragmentation. Each protocol has historically needed its own network of fillers. In 2024, Uniswap Labs and Across proposed ERC-7683, "Cross Chain Intents," to standardize how orders are presented to solvers. The current text, still a Draft created in April 2024, defines a solver-facing interface. Protocols expose orders through a resolver contract that translates them into common steps, payments and assumptions, so one solver can evaluate orders from many protocols without a custom integration for each. It deliberately avoids requiring a shared escrow or settlement contract.

The limits of chain abstraction today

Chain abstraction is real but partial. As of October 2026, these gaps are worth keeping in mind:

1. You still start on a real chain. Unless a smart account or gas sponsor covers it, the first transaction still needs that chain's fee token.
2. Coverage depends on solver inventory. Major pairs get tight quotes. Long-tail assets and newer chains may get wide spreads or no quote at all.
3. Risk is layered, not removed. An abstracted swap may touch a settlement contract, an oracle, a bridge and a solver. You don't see them, but each one can fail.
4. Refund mechanics vary. Deposit-address flows expect the exact amount on the correct network. Zodl's guide to NEAR Intents swaps, for instance, notes that a short deposit needs a follow-up transaction or is refunded after the swap expires.
5. Accounts can lock you in. A unified balance inside one provider's smart-account system is only as portable as that provider's chain coverage.

Where an aggregator fits

No single intent network covers every chain and pair at the best price. A bridge aggregator sits one level above. It asks several networks for quotes, compares what you'd actually receive, and hands you one deposit instruction. TorrentSwap does this across NEAR Intents, Chainflip and Flashnet, covering 39 assets on 17 networks. Every fee is already included in the quote, and it never takes custody: funds move from your wallet to the chosen network's settlement and back to your address.

Before you sign any abstracted swap, check four things: the destination address, the minimum you'll receive, the refund address, and roughly how long the route expects to take.

Want to see chain abstraction from the user's side? Open the TorrentSwap bridge, pick any two assets, and compare live quotes from several intent and liquidity networks. No account or signup is needed.