MPC: The Missing Distribution Link | Raghav Sood, Vultisig | ETHTaipei 2025

Hey everyone, nice to be here and sharing this with all of you. So I'm Raghar. I'm here to talk about MBC technology and how it can be used to address some of the current problems we have in user adoption, scaling this up, as well as introducing users to different parts of the crypto ecosystem and letting them maximize how they interact with the space. So, Vultisig is a at its core MPC wallet. It supports the majority of the top 30 blockchains by market cap.

It's been around for 1 year plus. Um, and it does a pretty decent job of being a wallet. But we're not going to focus too much on vault sync today, but instead on the underlying MPC technology and see how that enables us to address some of the aspects of three core problems that are limiting crypto adoption today. So if we look at crypto adoption right now, if you've been in the space and interacted a lot with retail or external users, it very quickly becomes obvious that a lot of people try to get into crypto and get pushed out at one of three different levels. So the first of course is just plain wallet setup.

It's difficult for users to have a reliable safe wallet setup experience right now. There is a lot of fragmentation with different wallets for different ecosystems, different ways to manage and store your seed phrases. Whether you're using a hardware wallet or a software wallet or a custodial wallet, there's a lot of complexity and hidden variability over here. Even if a user does manage to install a wallet, it's often not a good fit to allow them to explore crypto at large. If you start with an EVM chain, you're probably going to end up with MetaMask or Rabi or one of the EVM heavy wallets.

But as you start to explore other parts of the ecosystem, say you want to go to Bitcoin, you quickly end up installing multiple wallets, which further increases the burden of backing up multiple seed phrases and keeping track of your assets spread out across all of these wallets. Lastly, even if you do manage to enter an ecosystem, coming into crypto with no outside knowledge and limited support makes it difficult for users to discover protocols, interactions, uh places to deploy capital, tokens to work with. And this makes them susceptible to falling for scams or having suboptimal experiences simply because they don't realize there is a better offering out there and are unable to capitalize on all the great work that people have been doing in crypto. So if we look a little bit deeper at the seed phrase problem, it really just amounts to the fact that humans aren't wired for seed phrases. In any other scenario, almost everything that's important in your life can be replaced.

If you lose your passport, you can get a new one. You lose your driver's license, they'll print you a new one. You lose your university degree, you can send them an email and ask them to send you a new one. Almost everything is replaceable. But if you lose your seed phrase, you're toast.

There's just no way to recover from that scenario. And this means that people often rely on various backup methods. So you have extreme cases where people engrave it into steel plates and bury it in a safe. But a lot of people just use files held in a Google Drive or cloud storage with the assumption that they have more durability than what they would be able to manage as an individual. But this opens them up to a wide variety of attacks because now anyone who is able to discover this seed phrase has permanent unilateral unfiltered access to your funds.

It's difficult for users to split up key shares uh seed phrases and store them separately. There do exist schemes for this. You have shammer secret key sharing and all of that stuff but it's not approachable to the average user. Lastly, this often forces users to rely on centralized custody. A lot of people who enter crypto after viewing a headline like Bitcoin hitting a new all-time high simply end up leaving their funds on the centralized platform that they purchased for.

This is bad for the space. It's bad for users. It increases our systemic risk and reliance on these custodians. and it prevents people from engaging onchain in the way that we would want them to and get uh exposure to the innovations that crypto has. Similarly, cross-chain complexity means that users starting in one ecosystem often end up being overly indexed on specific wallets.

This again limits your engagement with not just that ecosystem but crypto as a whole because you will only be able to interact with chains supported by your first touch point. And if you want to migrate into other chains or ecosystems or even use certain features of your native ecosystem, you may end up having to install multiple wallets. We've already seen waves of this occur with a lot of people moving to wallets like Rabi from Metam Mask simply because they do a more comprehensive job of supporting protocols, indexing, EVM chains, RPCs. This creates a lot of fragmentation and it means that a lot of people tend to gravitate towards solutions that are narrowly scoped. So things like account abstraction are great, but they only solve one aspect of the problem and it means that users can get locked out of other parts of the ecosystem that haven't kept up with things like account abstraction.

And even if you do manage to get through all of that, you figure out your five wallets, you manage to touch point with everything that can happen in an ecosystem, you have an accessibility crisis. Protocol discovery is difficult. A user entering crypto may just end up on a or compound because they're old protocols. They're wellknown. They're easy to find without realizing that they could get better yield if they were to deploy assets through something like Morpho or Pendle, which would allow them to trade over and above the base a yields.

This means that a lot of users are limited in their utilization of crypto by their interface to crypto. It's difficult for them to discover and then reliably understand and evaluate new protocols which leads to users getting scammed which leads to users getting hit by ME attacks. It leads to users falling for fishing attacks and just makes the general on boarding experience for crypto significantly more unpleasant than it should be and often turns people off. If you get scammed in your first three weeks in a new space, there's a very good chance you're just going to leave and not continue exploring it. MPC tries to address these problems through multiple vectors.

So, how we're approaching this at Vultig is to approach it from a first principles basis. We want a solution that lets users start on the right footing. What this means is you should be able to set up one wallet, have it really, really secure and able to and transact on all of the different chains that you care about. This means we don't want to build upon chain specific or ecosystem specific solutions but instead drop down to the base cryptographic layer which allows us to generate a cross-chain multiasset wallet that is agnostic to the protocol and type of transactions that you're making. This means that as a user you benefit from distributed security.

You have multiple devices with multiple key shares. The compromise of any individual device does not result in any loss of funds. You have the option for using this to bootstrap programmable and authorization frameworks which allows you to install automations to help you make transactions across multiple protocols. And you have the option to rotate keys, which is most important. Which means that if you ever suspect that any of your devices has been compromised, you can rotate keys for the entire key signing committee and generate a fresh key set and discard the old one.

As long as it remained safe, you will be fine. A basic MPC setup will essentially look something like this, where you set up at least two or more key shares. So in this example we have three of these three we define a signing cororum of two. So this means if you install Visig right now you have your laptop and two phones you can set up a wault that looks like this and designate that any two of the devices have to be available to produce a valid signed transaction. This means you could lock one device in a safe.

You don't have to have it as a day-to-day carry. You can back up all three key shares separately as long as they're not kept in the same place. The compromise of any individual key share does not impact your security posture. This means that users avoid single points of failure when on boarding. They're able to work with flexible recovery options.

They get progressive security. You can start with a two of two and upgrade to a two of three or three of five without changing your addresses at down the road if you feel that you need more security. You get better recovery options which we'll touch upon in a minute and you get a more cross-chain single wallet experience which means you don't have to expose yourself to multiple providers, multiple implementations and multiple backup strategies. If we look at how this means for recovery, this lets you create a trust minimized social recovery system. Now, social recovery systems aren't something new.

A lot of account abstraction based wallets or even smart contract wallets like safe have had these for a while now and they basically rely during setup on you designating a trusted set of approvers who would be able to change the access routes to your contract wallet. Unfortunately, this only works for contract wallets or wallets which have a swappable authorization path underneath them. With MPC, you can apply this crosschain. You can apply it to Bitcoin, to EVM, to Salana, to Near, to Telegram, any primitive crypto ecosystem. And this means I could set up a wault, pick three people from the crowd here and designate each of them to be a recovery signer and use my two of three wault to generate a new committee that is a three of three.

That means as long as each individual in the recovery committee agrees to not collude with both of the other individuals, your funds remain secure. And in a disaster scenario, you can contact them out of band, work with them to recover access to your funds and create a new set of two or three key shares for your day-to-day use. The interesting part here is that as long as a single member of this recovery committee remains honest, your funds are safe. They can never unilaterally or even with a suboptimal collusion ever access your funds. This means if you designate a law firm [snorts] or a more trusted counterparty as at least one of the recovery members, you dramatically increase the safety of your funds and the security of your wallet.

This gives us a universal security model which allows us to define secure primitives that work across all different blockchains supported by our MPC cryptographic protocols. It's relatively futurep proof similar to your own vault. You can upgrade, rotate, and revoke access to your recovery committee. You can have multiple recovery committees. You can have committees that are time locked or timebased and have many other ways to stack and compose this.

This also opens up a more interesting use case which is a trustless agent installation. In a regular agent flow, when a user wants to use something like a sniper bot, they will typically open a Telegram chat, start a conversation, the bot is going to generate a private key, give them an address, and they hand over control of their funds, and essentially just trust that the bot is going to give it back. There have been cases where bots have rugged people. There have been cases where the sniperbot platforms themselves are compromised and both have resulted in users losing funds. With an MPC based setup, we can use our base wallet, our own personal access, which is the two of three key shares to reshare into a bot or a genic committee.

And this will result in the creation of two new key shares, an agent key share and a guardian key share. This is a two of two committee. So both the agent and the guardian have to cooperate to sign any transactions. Both of these key shares are held by two separate entities. The agent proposes a transaction and the guardian agent verifies that it is a valid, safe, legitimate transaction.

If the guardian refuses to sign, there is no way for the agent itself to unilaterally ever move funds. If you want to revoke access to the agent, you simply ask the guardian to delete its corresponding key share and you have a cryptographically sound proof that the agent can never transact from that wault again. This lets us create a lot of interesting bolts because now you have a scalable way for other automated platforms to propose transactions on your wallet and ensure that they can be autonomously or manually approved depending on how you set up your key share committees. So, and allow transactions to be done uh out of ban. So, you no longer have to give up custody, nor do you have to be available 24/7 to approve a wallet connect popup, but you can instruct a bot and say, "Hey, help me DCA $100 a day into Bitcoin for the next 30 days."

And as long as the Guardian server is aware of that mandate, it will be able to approve transactions. But if the bot tries to do something else, maybe buy a different coin or make a withdrawal, it will refuse to cosign. This means that policies are enforced before signing, which means at no point does a valid signed transaction exist that could be harmful to you versus some other systems which enforce policies based on post signature simulations. This means you can scale up an MPC wallet to essentially create a one interface for every protocol to address the D5 complexity problem. Right now, a lot of users may not be able to optimate uh optimize and access protocols like Pendle or Moro or many others simply because they're too complicated.

Rates trading is not something that the average user does daytoday in their web to traditional finance life and coming to crypto and having the expectation that people will do this is a little bit lofty. This means protocols, wa aggregators and operators can instead create bots and agents which users can trustlessly install to their wallets which help them optimize yield. They would simply be able to talk to their wallet and say, "Hey, I have 10,000 USDC. Find me a lowrisk minimum 5% yield on this over the next 6 months." And the bot will then say, "Okay, I can make initial deposit into a then go trade the PT and YT tokens on Pendle and give you a fixed rate for 5% at certain risk parameters."

And this is something that is unique to the MBC world. It's very very difficult to replicate this outside of an MPC setup. This means you can have trading agents that do automated trades on Texas. You can have limit orders created even on AMMs that don't support them because an agent can monitor the price for you and automatically propose and execute a transaction when your limit prices are hit. You can have agents that protect you from scams.

So you could have an agent that says, "Hey, every time I make a swap, check to see if the token contract matches the token contract listed on a verified source like Coin Gecko." And if it doesn't, refuse to sign it. You can have agents that help you navigate and discover opportunities. So instead of saying go to a and deposit into Morpho after that you can just say help me find yield and the agent will be able to access its index and onchain data for various protocols and come up with a good yield strategy for your assets and your risk profile. You could have auto rebalancing.

You could run your own index fund entirely self-custodized without trusting any external parties. You could have riskbounded execution, which could be things like capital preservation agents, where if they detect a sharp market drop while you're asleep, they can help you exit risky positions without you having to constantly monitor the market. All of this can be built as agents and bots and actually offered as a service to users through a marketplace-like experience which allows developers to essentially bundle execution strategies as automations and sell and offer them to users for free or for a subscription. This means that MPC is great for users. It enforces self-custody.

It enforces safety. It allows users to use their existing devices. There's no need to buy specialized hardware wallets or any special tokens. You have users that can rely on recovery methods they're already familiar with. Family, friends, legal entities.

There's limited technical knowledge needed. A good MPC wallet can abstract away the majority of this behind the scenes. And it's one wallet for all chains. As a user, you don't have to constantly manage seed phrases, multiple seed phrases, their backups, and how to work with them. MBC is great for builders.

It allows builders to have a cross-chain single entry point for transaction signing and proposal. It allows builders to access funds held in a non-custodial manner without taking on the custody risk and all of the overhead that comes with it. As a builder, you can build bots and focus on your execution and transaction proposal without having to worry about whether you're going to get hacked and accidentally leak someone's private keys. It's great for developer experience. You're not bound to the availability of certain EIPs or certain implementations of account abstraction on every chain that you want to work with.

As long as you can work with the basic primitive crypto curves, you will be able to work with any chain that uses them. So this means you can come up with execution flows that are crosschain from day one without having to refactor to support multiple protocols or implementations of wallet systems. It has built-in security guarantees. You have proof that if a user uninstalls your agent, you will never accidentally continue to use their funds. There is no leftover or stale approval concept here.

And you have policy as code because you can instruct verifiers to effectively evaluate your proposals against a predefined policy which can prevent bugs or other issues from slipping through. So this brings us down to Walty. Walt, as I mentioned, is a wallet. We're focused on not allowing users to fall for common traps on security like seed phrases. We support thresholds, multiple devices, all of that good stuff.

It's based around DKLS23, which is a newer and significantly more performant MBC protocol, allowing users to have fairly efficient and often faster than hardware wallet, key sign, and key gen times. Everything we do is open-source and audited. We don't want to build in private. We want to build in public. And we want to make sure that we're giving the best platform to people to build upon.

We have native applications for Android, iOS, Linux, Mac OS, and Windows. It's all native. No shared codebase. No uh React, JavaScript trickery here. Everything is done natively.

What this allows people to do is have a unified problem uh platform for solving fundamental problems. Users get to keep their assets safe and builders get access to these user assets as deployable TVL to route to their own protocols and extract yield from the future that we're looking at is effectively eliminating seed phrases. We're already here. We have a wallet. It's been live for 1 year plus.

It works great. Our long-term vision is to build this into an infrastructure which allows builders to come and generate their own apps, bots, agents, and automations similar to what platforms have done with account abstraction and safe wallets and allowing these people to access the funds in various user wallets trustlessly. Oops. Uh, okay. I guess I think I might have Oh, okay.

Yeah, I guess we're back. Um, yeah, I'm looking forward to seeing more people building on MBC and crosschain solutions. Happy to take any questions here or over Telegram or if you want to find me after, that's great, too. Any questions for today? Yes.

Yes. Thank you for your sharing. I'm just confused about the you you mentioned the one wallet for the all chains. It's it you mean the all chains mean the EVM compatible chains or uh like the Sonana? No, I mean all chains.

So for multi-IG and for MPC in general, you usually need two implementations of the protocol. One for ECDSA and one for EDDDSA. Yes, ECDSA covers EVM, Bitcoin, Litecoin, Dogecoin, all of your SEC 256K1 chains. While EDDDSA covers Solana, near to many of the newer ecosystems. Uh most MPC protocols today have implementations for both including Vultisig.

So this means a user using Vultig can access Bitcoin, EVM chains, Solana, Toron, Ripple, Near um almost every major chain in the top 30 50 by market cap. The only chains that we are missing at the moment are some of the more esoteric cryptographic ones like Monero. But this is a work in progress as well and should be live later this year. But when we say all chains, we mean all chains, not just EVM. Okay.

So you mean one W means the same W address for the all chains. So it works similar to most other multi-chain wallets. We rely on BIP 32 3944 for crosschain multi-chain address derivation. That means you only need to back up your key shares once and that contains the root keys for all the different [laughter] chains. Okay.

Okay. Interesting. Thank you. Any other questions from the audience? Oh, there are many other MP solutions like uh fire firebox firebox or coinbase some uh recently open sourced MCP protocol.

What's the difference between uh them and benefit of barik? Yeah, I think that's a great question. So, there are a few different ways to implement MPC. Probably one of the more well-known ways is the GG20 implementation which Binance open- sourced a solution for about almost 10 years ago now and that has been used very commonly in a lot of different applications. The downside to GG20 and many of the older solutions is that key gen and key sign are slow which means it can take you 15 to 20 seconds to sign a transaction and maybe one or two minutes to generate a new wallet.

We picked DKLS 23 specifically because it's optimized for faster actions on the cryptographic curves that are relevant for blockchains. This means we can get keygen down to 3 to 5 seconds and key sign in many cases to under 1 second. Uh in reality it takes a little bit longer because of network costs when communicating between devices. Uh but this gives you a very very good experience as a crypto user because you can actually transact even on very fast chains without running into issues like block time expiry or swap code expiry. If you compare it to Fireblocks, uh that's a little bit new nuanced.

Fireblocks in particular uses something called MPC cgmp. And the main benefit of cgmp is that it lets you do outofband signature aggregation. So if you're using Voltisk or Coinbase's MPC implementation, both of the signers have to be available and connected to each other at the same time. While if any of you have used Fireblocks, you will know that you can actually have people add their signatures 20, 30 minutes, 2 hours after each other. They don't have to be doing it at the same time.

Uh this has benefits. Namely, you don't need everyone available at the same time. But again, it comes at the cost of speed and it comes at the cost of complexity because it makes it a much slower and more expensive protocol to run. Coinbase's MPC is a little bit more comparable to Walty, but it's also slower. Um, but at the end of the day, it does have a lot of the same benefits, and it is something worth considering if you're looking to build on an open-source MPC platform.

Uh, both Coinbase and Wyig will be a good fit. Thank you. Any more questions? Okay, thanks so much. If anyone has any other questions, feel free to find me.

I'll be wearing a Valty Sig hat after this. I should be easy to find. Thanks for attending. Thank you, Rahav. [music]