A special thanks to Arjun from Li.Fi, Jay from Eco, Max from Everclear, Pablo and Margaret from Bootnode, Mike from Nomial, and Connor, and Nam for their feedback and review.
Key Takeaways
- Intents and message-based bridges offer two distinct architectures for transferring assets across chains, each with their own set of tradeoffs.
- Message-based bridges are designed to be reliable and secure across many different chains, but their performance is bound by origin chain finality times.
- Intent-based bridges are designed to be fast and convenient for users, but face more difficulty with supporting larger transfers and accommodating for a greater number of chains.
- When transferring value across chains, intents provide a smoother experience for retail users, while messaging provides greater reliability and security.
- The two architectures are complementary to each other as message-based bridges support critical backend operations for intent-based bridges.
The Interoperability Landscape Today
As the number of blockchains continues to grow by the day, interoperability has become a key consideration for builders and users alike. Builders — whether its application developers, chain deployers, or asset issuers — need to think about how to make their product extensible across multiple chains and ecosystems. Meanwhile users must figure out how to navigate across different chains and ecosystems, whether it’s to buy the latest memecoin on Solana, play the latest game on Avalanche, mint the latest NFT on Zora, and plenty more.
To date, however, token bridging is by far the most common use case for blockchain interoperability. A user has funds on chain A, but wants to get those funds on chain B. To do so, they need to use a bridge.
Bridges come in different shapes and sizes, but the key difference lies in how they enable tokens to be transferred across chains. The two key architectures present today are message-based bridges, and more recently, intent-based bridges.
In traditional crypto twitter spirit, there’s been a number of semantics debates around what intents really are, and whether one design approach is better than the other — “Team Intents” vs “Team Messaging”.
However, a closer look at the architecture of intent-based bridges reveal that intents and messaging are NOT mutually exclusive architectures. In fact, it’s becoming increasingly clear that the two are complementary to one another.
Understanding Message-Based Bridges And Intents
A Primer on Message-Based Bridges
Message-based bridges enable data to be transferred across multiple chains. Arbitrary data, such as token transfers, contract calls, price feeds, and governance votes can be sent across chains using predefined smart contracts, similar to using an API. Message-based bridges have been a primitive in blockchain interoperability since the advent of the Cosmos IBC protocol, which itself drew inspiration from the TCP/IP specification that laid the foundation for the modern internet.
Message-based bridges explicitly determine how tokens will be sourced and delivered ahead of time, by minting representative tokens through a smart contract on the destination chain. Token standards like Warp Tokens play a critical role in this, as they offer predictable outcomes by locking or burning tokens on the origin chain to mint an equivalent amount on the destination chain. This approach provides a reliable source of liquidity, eliminates slippage, and is adaptable to handle transfers of any size.
How Do They Work?
Transferring tokens between chains using a message-based bridge involves two transactions: one on the origin chain to initiate the transfer, and another on the destination chain to complete it. When a user initiates a transaction, a message is first sent to the designated smart contract on the origin chain to ensure message authenticity. The message is then picked up and transported to the designated destination chain smart contract, where it is verified and then delivered to its recipient.
Most message-based bridges follow a similar structure, but come with their own implementation details. For instance, on Hyperlane, messaging smart contracts are called Mailboxes, which are currently live on 85 chains. Validators attest to the authenticity of outgoing messages on the origin chain. Security is facilitated on the destination chain by smart contracts called Interchain Security Modules (ISMs), while Relayers help transport and deliver messages across chains.
Messaging Bridges: Bound By Finality
As mentioned above, message-based bridges require two transactions to transfer tokens across chains. This means the performance of messaging bridges is bound to the design of the chains they connect.
Before a message can be delivered on its destination chain, the bridge may need to account for the finality of the origin chain. A simple way to think about finality risk is the “uncertainty that a transaction, once completed, may not be as final as it appears” — Across Blog.
In more technical terms, finality defines the point at which a transaction is permanently included in a chain and it is safe from a blockchain reorg. Finality occurs after a certain number of blocks have been added following a particular transaction, which can greatly vary by chain.
It’s worth noting users making larger transfers across chains value reliability and security over delivery time. People still continue to use rollup bridges today, which can take as long as 15–30 minutes, to transfer tokens across chains. However, many users simply want to be able to bridge assets across chains as seamlessly as making a single-chain swap on a DEX. A $10 transfer doesn’t quite require the same security assurances as a $10,000 transfer.
This is where intent-based bridges come in.
Understanding Intent-Based Bridges
Intent-based protocols are based on a simple concept: user transactions are executed by third-party agents on their behalf. A user expresses their intent, and a specialized agent — commonly known as a Filler — completes it for them. Fillers provide the necessary capital upfront to fulfill user intents immediately, and are later reimbursed by the protocol through a settlement process.
Cross-Chain Intents
Intent-based protocols started out serving single-chain swaps, pioneered by CoWSwap and 0x. However, a new generation of protocols — pioneered by Across — saw the potential for intents to power transfers across different chains, and thus intent-based bridging was born.
Instead of directly sending a message across chains to transfer tokens, an intent-based bridge outsources the token transfer to a filler. The filler is then responsible for delivering the user’s tokens on the destination chain, and is repaid by the bridge at a later time through a settlement process.
How Intents Work
The lifecycle of an intent can be distilled into three key stages:
1) Expression, 2) Execution, and 3) Settlement.
Users are responsible for the expression of an intent, while the protocol and its fillers handle execution and settlement.
- 🗣️ Expression — A user expresses their intent by signing an offchain order. (i.e. “Swap 100 USDC on Ethereum for 100 deUSD on Sei”)
- ✅ Execution — A designated filler fills the user’s intent (Filler sends 100 deUSD to the user on the destination chain).
- 🤝 Settlement — The filler submits a proof that they did in fact deliver 100 deUSD to the user on Sei. The user’s funds are then released from the settlement contract and delivered to the filler.
It’s important to note that user funds first get sent to the underlying protocol to be held in escrow, and are released to the filler after the protocol has verified the fill.
Where Intents Shine
From a user perspective, the key difference between messaging and intent-based bridges lies in speed and delivery times. While both approaches require two transactions to transfer tokens — one to initiate and one to fulfill — intent-based bridges don’t need to account for finality on the origin chain in order to deliver a user’s tokens on the destination chain: instead, their fillers do.
To fill an intent, fillers loan out capital to deliver tokens on the destination chain right away. By doing so, they isolate finality risk to themselves, as they assume the risk of not getting repaid for their services in the event of a blockchain reorg. Importantly, there is no inherent speed limit for fillers, as they can always choose to fill an intent or not, and can fill intents as quickly as they are comfortable assuming finality risk. However, given that this design allows users to enjoy near-instant transfers, fillers are incentivized to fill intents quickly and charge users fees accordingly.
Liquidity sourcing highlights another way intent-based bridges set themselves apart. While messaging bridges rely on frameworks like Warp Tokens to mint and deliver representative tokens to users on the destination chain, intent-based bridges let fillers choose where to source liquidity from. Depending on the protocol, fillers can use personal inventory on the destination chain, swap for the token through liquidity pools, or even use offchain liquidity on CEXs.
The takeaway here is that intent-based bridges don’t need to rely on a contract for minting tokens, but can instead work with any asset type, including assets secured by native bridges, often referred to as canonical assets.
Where Intents Fall Short
However, intent-based bridges are not without their limitations. For starters, these protocols rely on the availability of fillers over a long period of time. If there are no fillers willing to fulfill an intent, users have to bridge their tokens elsewhere.
Bootstrapping filler networks, however is no easy task. Fillers make money by charging fees for filling intents and without meaningful order flow passing through the protocol on a regular basis, there is little incentive to dedicate time and resources towards that protocol.
To further compound this issue, barriers to entry, such as high costs and operational complexity, make it challenging to participate as a filler in a filler network today, leading to a small number of fillers dominate a majority of order flow.
As Arjun Chand highlights in “With Intents It’s Solvers All The Way Down”, this dynamic introduces monopoly risks (i.e. extractive fees charged to users, order manipulation by fillers) as well as liveness risk in the event that a major filler goes down or is unable to operate further. Emerging standards like ERC-7683 aim to level the playing field for fillers by standardizing intent formats and imposing mechanisms such as refund guarantees and filler data submission.
Scalability is also a key concern with intent-based bridges. Intent-based bridges require fillers to have capital available on the chains they serve, which becomes increasingly difficult to support at higher transaction sizes and a greater number of chains. It’s easy for fillers to maintain capital on 2–3 chains, but what about 100s of them? And what if someone is transferring millions across those chains?
It’s worth mentioning that a number of teams are working on solutions for optimizing liquidity management for fillers. For instance, Nomial is building liquidity pools that fillers can borrow from to fill intents, eliminating the need to hold inventory across many different chains. Everclear just recently introduced a vote-bonding system (vbCLEAR) to enable chains to lock tokens and allocate incentives to fillers to support their ecosystems. These solutions are expected to become critical in helping decentralize filler networks at scale across multiple chains and ecosystems as the intents landscape grows.
The Convergence of Intents and Messaging
It’s become evident that of today, intent-based bridges are designed to optimize for speed, but face limitations around high bridging volumes, whereas message-based bridges optimize for reliability but are limited by speed. So does this make one bridging architecture ultimately better than the other?
Not quite.
While intent-based bridges and messaging bridges implement different approaches for enabling cross-chain token transfers, they are quite complementary, as messaging bridges support several key operations of intent-based bridges:
Settlement
Settlement in intent-based protocols is similar to how Visa operates, as Arjun Bhuptani describes below:
In intent-based bridges, fillers get reimbursed by the protocol for fulfilling a user intent, receiving the amount they paid out plus any associated service fees. Fillers must submit a proof of their transaction to the underlying protocol for verification, and once verified, they can receive their funds. This process involves sending a message back to the origin chain where the user expressed their intent, which can be done using an oracle like UMA’s Optimistic Oracle, or a message-based bridge like Hyperlane.
Message-based bridges provide flexibility for protocols and fillers to choose their preferred verification mechanism. For instance, with Hyperlane ISMs, verification mechanisms can be as simple or complex as needed, and can easily be swapped out for new mechanisms as they come out.
Rebalancing
After filling an intent, fillers often need to rebalance their inventory across the chains they support to ensure liquidity is available for future fills. Message-based bridges play a crucial role here, enabling fillers to rebalance large volumes without being constrained by capital limitations, as they can mint new tokens on demand rather than holding a pre-stocked inventory. This approach is especially useful for serving newer, long-tail chains where liquidity and user activity tends to be more scarce.
Furthermore, message-based bridges are significantly faster than native rollup bridges, which can take up to 7 days for L2<>L1 transfers. As such, message-based bridges provide a faster and more accessible way for fillers to rebalance inventory across chains.
One noteworthy solution to this Everclear’s Clearing Layer, which aims to streamline settlement and rebalancing for fillers. Rather than having fillers manage multiple separate rebalancing transactions for each chain, Everclear nets the total inflows and outflows across all supported chains, using Hyperlane as its underlying transport layer.
The Takeaway: Intents 🤝 Messages
Intent-based bridges have been slowly but surely eating away at total bridging market share over the last year. Due to the improved experience and isolated finality risk they offer, some believe that intent-based bridges may even end up overtaking message-based bridges eventually.
This perspective, however, overlooks how complementary intents and messaging really are, even if the majority of users flock towards intent-based bridges in the future. As the number of intent-based bridges grows, so does the importance of message-based bridges. Whether it’s providing efficient settlement for fillers or lowering their rebalancing costs, message-based bridges provide the foundational support intent-based bridges need to scale and grow over time.
Intents and fillers facilitate the user experience, while messaging bridges facilitate the filler experience.
Intents, Powered By Hyperlane
Hyperlane is an open permissionless framework for passing messages across different chains. Its modular nature allows it to be configurable to a variety of use cases, including intents settlement and rebalancing. Today, a number of teams are building intent solutions powered by Hyperlane:
- Skip Go Fast is a decentralized bridging protocol enabling users to bridge from Ethereum and EVM rollups to the Cosmos ecosystem with one click, using Hyperlane for settlement. Users express intents on the origin chain, and fillers help deliver their funds to the destination chain, allowing users to bridge to Cosmos chains in a matter of seconds. Fillers then settle with the Skip Go protocol afterwards, using Hyperlane to submit confirmation messages from the destination to origin chain. Read more here.
- Eco Protocol enables fast, cheap, single-click stablecoin transaction execution anywhere across Ethereum. Using Hyperlane, Eco was able to implement batch settling, a transfer method which spreads transaction costs across multiple fillers, making each transaction significantly cheaper.
- The Compact is an ownerless ERC-6909 contract that allows token holders to create reusable resource locks. In short, Uniswap users sign an intent for a cross-chain swap that commits to a Hyperlane arbiter contract for filler settlement. As The Compact is still in development, the latest progress can be observed in the Github repo: https://github.com/Uniswap/the-compact.
What’s Next?
Hyperlane is built on the premise of permissionless interoperability: anyone should be able to connect any chain, on their own terms. The intents landscape should be no different.
Demand for intents has been steadily growing over the past year. However, building intent-based protocols is not an easy task by any means today: builders have to account for bootstrapping a filler network, configuring the right intent <> filler matching mechanisms, rebalancing and settlement logic, UI templates, and plenty more.
These challenges have demonstrated the need for an open framework to make building intents-based protocols simple and accessible for everyone.
This is why Hyperlane contributors, alongside Bootnode and several other teams, have started assembling a working group to bring together the worlds of intents and messaging. Anyone interested is highly encouraged to reach out in #developer-intents-framework in the Hyperlane discord and get involved!
More about Hyperlane
Hyperlane is the open interoperability framework. It empowers developers to connect anywhere onchain and build applications that can easily and securely communicate between multiple blockchains. Importantly, Hyperlane is fully open-source and always permissionless to build with.
