# Andrei Duma - The Death of the Multi-Step Transaction

- Channel: [ETHCluj Meetup](https://streameth.org/ethcluj-meetup)
- Date: 2026-06-09
- Duration: 16:50
- Watch: https://streameth.org/watch/yt-SM1JYWLSD60
- YouTube: https://www.youtube.com/watch?v=SM1JYWLSD60

## Description

A talk about how fragmentation became DeFi's biggest UX enemy and what it took to rethink deposits as a first-class primitive. Deposits in DeFi have always been a complex, multi-step process requiring swaps, approvals, bridges, and multiple protocol interactions, often across chains. This talk makes the case that simplifying execution is as important as building new financial primitives.

## Transcript

Hello everyone. Good afternoon. Just to introduce myself briefly, I'm the head of DeFi at LiFi or LiFi. Technically, it's LiFi, but uh it's just weird to say DeFi Lei, so I I I pronounce it as LiFi. Today we're going to talk about uh multi-step transactions in DeFi and how that is a huge pain and how we at LiFi were trying to solve this. Uh this presentation actually has been built for a 30 minute slot but um I'll try to you know wrap everything in uh in the 20 minute allocated. So to kick things off four years ago we used to call deposits zaps. That's how everyone in DeFi understood deposits. A zap was basically a swap bridge swap and a deposit. However, nowadays it's just purely deposits. That's how kind of the uh the users in DeFi under are understanding this. The problem is that liquidity is not only very fragmented in the D5 landscape. It's all it also comes down to the liquidity pools and the contract addresses and how you want to do the actions on chain because right now and there's actually a zero here if you want to do a deposit you you need to do in some cases a swap on the source chain then you need to bridge the funds. So that is basically one signature. Then you have to approve the token. Then you need to swap your asset on the destination chain. Approve the vault deposit and then you know check on chain if it worked. So it's it's quite a pain. And for any anyone who did uh a deposit into any vault, you probably know that you don't have the source asset which might be you know ETH or USDC because you know why would you hold ETH in uh in DeFi nowadays. Uh so we we're trying to abstract everything into a one-click experience and we know that user retention actually has a big drop off because it's complicated and because you need to do so many actions and it actually like we timed it. It takes 15 to 20 minutes to do a crosschain deposit into a vault and we're just talking like purely vaults for earn. Uh to go a bit back in the history, initially we just had the transfer assets. So Bitcoin, Bitcoin, ETH, ETH, then we had the swaps, then we had the bridging and nowadays anything can be combined into a deposit. The deposit basically started after DeFi summer in 2020 when people realized that we don't have yield on chain. So when yield came on chain then we obviously had a new problem that we we had to solve. And why was this never a primitive? Because yield wasn't designed to be on chain. Yield on DeFi uh on Bitcoin on Ethereum on USDC. And you the problem is right now that you have so many different types of contracts. So nothing is unified. Then you have the crosschain overhead where where whereas you have different types of bridges, different types of intentbased dexes and there's no shared execution layer. So basically you cannot go from ETH to USDC on a for example as a destination chain to get your yield there. You have to do as I said you know plus six transactions. But we've uh we've worked on this and we actually as I said we had zaps you know 3 four years ago but it was a very uh bad sort to say experience even for the users back then because zaps 3 years ago were basically designed through certain bridge bridges which had what we call destination calls. So a destination call basically allows you to do uh to call the contract after you breached in the same transaction and do the deposit. The biggest problem was that there wasn't any uh simulations. You also had a lot of dust in the wallet that remained. So we've worked on this for 24 months and we built something called composer. So composer basically allows you to abstract everything into a oneclick experience because you have an explicit an explicit step composition. So the developers declare what the actions one are before they deposit and then you have to also validate uh before signing you have the dynamic call data call data injection. So basically you take the output of any transaction and you can use it as an input to roll over into the next transaction and I'll I'll speak about it in a few seconds that can be chained onto multiple transactions. So deposits is just the first step, but you can do so much more because now you have a primitive that allows you to have a chain of events like swap bridge swap deposit and then uh take a loan and then put a limit order as a stop-loss for example for your position. And I think it's very important to mention also that you have lower lower gas overhead because you're batching multiple transactions into one. So you don't have to basically click six times or you know pay for the gas multiple multiple times. Uh as I said I wanted to get here so to the V1 that we're shipping right now because in V0 it's any token on any chain to deposit. So any token on any chain to any action on any chain is what we just released right now at LiFi. It basically allows you to do any atomic flow. And in 80% of the cases indeed is going to be just deposits, just you know vault deposits. But vault deposits include looping sometimes. And doing looping in a vault necessitates a lot more than just the deposit because you do the deposit and then you take uh a loan against your deposited uh sum and then you loop that multiple times. So you basically increase the leverage. Now, if you're doing that, you obviously have to batch multiple transactions in the contract. And also, in some cases, you might want to have a stop-loss because you don't want to wait until your health ratio goes uh goes down. So, in this case, we take the we do the swap, we call the contract, we add the guardrails. The guardrails are basically what I just mentioned, which are what you want to do within the contract. Then you have the deposit into the vault and then you have the deposit into the next vault or the looping and then you can have a stop limit order and whatever you as a developer are inputting in the contract. And to showcase another example, we we just talked to a big market maker in in DeFi. And this is something that even market makers could use because they when they bridge funds on chain and they manage liquidity on chain in many cases you'd be surprised how many manual interactions are happening. So this basically makes your life a lot easier as a developer as a market maker as if you want to have an earn front end and uh so on. So, as I said, you know, six uh six steps, two signatures, way too many fails, way too many pain points to actually go get to the point of what I wanted to to say here is we productize this in a way so that neo banks nowadays can use composer because composer is basically a tool just for developers. part, but as a as a you know, just a normal DeFi user, you want to have uh access to a product and maybe you're not that DeFi say uh savvy. So that basically means uh LeFi earn what Leifi earn is it basically allows you to get yield on chain. It allows you to have the data part because composer is just the engine which is the most important part. However, you also need the data part. And for the data part, I mean you need to have the analytics, you need to have uh the historical APY, the projected APY. So if you combine those two together, you basically have Leifi earn which allows yield from anywhere on chain. Uh you have the full execution flow handle, you have the composer powered safety, and you can launch a yield product, not a stack. So you basically can choose whatever yield product you guys want and we we can just add it into composer and we're going to composer is basically you know a typescript so it's super easy to add new integrations with one click start to finish. So this basically is the productized version of composer. So you click deposit, we simulate obviously the full path. Then we have you know as an example east to arbitrum, east to USDC and then USDC enters the vault and allows users to get uh to get the yield on chain and sometimes there's compounding of the yield on chain. So you can also do the compounding within the contract through composer. Uh flows that used to take a quarter. Something very important that we we realize is the dust sweep. So users have a lot of dust, a lot of unnecessary tokens, small amounts of tokens in their wallets. So as a NEO bank or as a wallet because most wallet most wallets nowadays are becoming a NEO bank you want to allow your users to have certain features which centralized exchanges have. You probably know that on Binance or on any of the big centralized exchanges you have a dust sweep. So just sweep everything that you have as low amounts in your in your centralized exchange wallet. The problem is is not really a painoint that has been solved in DeFi. So with composer you can basically dust sweep everything and get it into your dedicated uh token like ETH for example or like USDC. On top of that you can do a dust sweep and a deposit. Let's say that you have multiple tokens on multiple chains and you want to swap bridge swap them and then take them on a s on a destination chain and deposit them into a new uh into a new contract. What we're also working on cuz we're we're very businessdriven. So when we hear feedback from our users and from our clients, that's what we're you know that's where we're heading is the multivol deposit. So you can split a single deposit across multiple vaults. This is something that we actually think is going to be a new primitive in DeFi because it allows people if you have a source token on a source chain, but you want to split your tokens into different yield opportunities on different chains. You just think about it. You you probably need four to six signatures and steps for one deposit. If you multiple if you if you multiply that by two or three then it's it's going to take you you know like a halftime uh day to actually do that. So you abstract that into everything. And also the portfolio rebalancing. The portfolio rebalancing allows composer you just give it the input. You're you're telling composer what you want composer to do which is basically rebalance at every 6 hours based on the APY if it falls under or if it's uh the targeted APY and then it rebalances everything into uh into the asset that you want. Everything that I spoke about is live in production already is being used. Uh so far we processed about 20 to30 million in TVL through composer and through the earn product is live on Jumper. Jumper for uh for the ones that don't know is our B2C product. It's jumper.xyz. It allows you to do swapping, bridging, getting yield on chain. It's a super app. It's actually the most used crosschain venue for aggregation right now in DeFi. I suggest you try it out. I suggest you go on jumper.xyz and actually uh try to do a swap, a bridge, and also for the earn part, you don't need to do the swap and the bridge beforehand because you can do it all into one transaction. Uh there what you don't have to build anymore. You basically don't have to do the hooks for the custom bridge and the swap. You don't have to do the app specific approval batching logic. You don't have to do the per chain per protocol deposit user user experience. You don't have the fragile multi-transaction state management. So all those here are as for you know the ones of you that are developers are huge pain points which took us 24 months to solve and to actually productize it and ship it for our users. So I think that with composer and with the lei earn product you solve for so many issues. So vault is just one of one of them and that's why we have lefi but if you want to do if you want to build on top of composer that's super easy and for any developer they can do that right now basically given what actions they want to do. So I think that you should remember this presentation as in you can now do any action on chain batched together into a one-click experience. swaps, bridging, deposits, limit orders, anything can be bashed into a one-click experience for your users and for the apps that you're building. Uh, capital flowing from anywhere to anywhere is basically, as I said, you know, you want to have a seamless experience for your users. making capital move effortlessly effortlessly uh is as important as building places uh to put it basically what I wanted to say through this is that you have capital in your wallet that's just sitting idle and in many cases you could just put it to work in a very effortlessly way I in my example I've been a power user for many many years now I didn't seek yield mostly because I knew that one I have to search for it. There's a lot of work in searching the yield and that and then there's also the execution part. So now basically I just use VFI earn to put my tokens to work. I'll leave you with uh one question to think about. If you could ship one multi-step flow next month without building it inhouse, what would it be? If you could do anything that requires multiple transactions, what would you ship in the next few months? It can be a swapping interface, it can be a bridging interface, it can be both. It can be a yield uh opportunity, it can be uh a limit order, it can be a per venue. Whatever would it be, think about it and just you know remember that you have composer as a tool for you as a developer and also lei earn for you as an end user.