New Ethereum talks, every Monday. The week's conference uploads by event, in your inbox.

Loading player…

Walrus Workshop I Daniel Leavitt

ETHGlobalThu, Oct 9, 2025, 12:00 AM

Daniel Leavitt, a Devril engineer at Mysten Labs, introduces Walrus, a decentralized storage network built to address the shortcomings of current blockchain and Web3 solutions, such as centralized frontend hosting and NFT metadata storage. Walrus leverages a novel encoding algorithm for cost-effective, scalable, and reliable data storage with a low replication factor, outperforming traditional decentralized networks. It uses the Sui blockchain for coordination, enabling programmable, composable storage solutions and tokenized storage capacity. Walrus is chain-agnostic, with public-by-default data, and supports integrations via TypeScript SDKs, HTTP APIs, and S3-compatible APIs. The workshop highlights developer tooling, cross-chain support, and app-building opportunities, including Walrus Sites for decentralized frontend hosting. Daniel discusses hackathon tracks, resources, and answers questions about cost, data sovereignty, and use cases, emphasizing Walrus’s strengths in censorship resistance and data ownership. 00:00 Introduction 00:31 About Daniel & Mysten Labs 00:58 Why Walrus Was Built 02:37 Mysten Labs’ Mission & Sui Blockchain 03:03 Walrus Architecture & Encoding Algorithm 05:01 Programmability & Sui Integration 06:14 Walrus Token & Decentralization 06:42 Data Privacy & Seal Product 07:36 Walrus Sites: Decentralized Frontends 08:47 Hackathon Track Overview & Prizes 09:11 Developer Tooling Track 10:31 Cross-Chain & ENS Support 10:53 Best App with Walrus Track 12:21 Developer Resources & SDKs 13:37 Chain-Agnostic Integration Methods 14:28 Partner & Community SDKs 15:07 Awesome Walrus & Community Tools 15:34 Live Demo: Uploading Data 16:25 Tips for Hackathon Participants 17:31 Support & Q&A 18:05 Q&A: Cost, Use Cases, Data Sovereignty 20:20 Walrus Sites Deployment & Portals 21:58 Favorite Use Cases & Future Opportunities 22:46 Storage Duration, Renewals, and Deletion 24:10 File Size Limits & Testnet Info 24:48 Closing Remarks _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 🇺🇸 *ETHGlobal New York* This workshop is specifically for ETHGlobal New York, a 3-day hackathon held August 15-17, bringing together the most skilled web3 developers, designers and product builders from all around the globe for a weekend-long adventure to advance the Ethereum ecosystem! Watch the full ETHGlobal New York 2025 playlist here: https://www.youtube.com/playlist?list=PLXzKMXK2aHh4Xze0ZeSWCSGsw25clPpvZ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ✅ *Follow Walrus* X: https://x.com/WalrusProtocol ✅ *Follow ETHGlobal* X: https://x.com/ETHGlobal​ Warpcast: https://warpcast.com/ethglobal Website: https://ethglobal.com YouTube: https://www.youtube.com/c/ETHGlobal _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _Are you interested in Ethereum development and entrepreneurship?_ 👉 Sign up for the next ETHGlobal event: https://ethglobal.com/events 🎁 Get exclusive access and perks with ETHGlobal Plus! https://ethglobal.com/plus 📣 Want us to throw an event in your city? Tell us where! https://ethglobal.com/city _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Transcript

[Music] I hope you're all having a good start to the hackathon. We'll just wait a few more people coming in. All right. Um, so I've probably talked to some of you, but for those of you who don't know, I'm Daniel. I'm one of the Devril engineers at Miston Labs.

Misten Labs, the company that built both Walrus and the Sweet Blockchain. So, I'm familiar with both. I'm able to answer questions from both. Um, yeah. And today I'm going to be explaining Walrus, uh, why we built it, what makes it better than other decentralized networks, why you should use it, and I'll go into what we're looking for in the track, and how you guys can get started.

So, first I'm going to talk about why we built Walrus in the first place. So, blockchains are meant to decentralize the web. You know, we're supposed to put ownership back in the hands of the users and everything. And for the most part, blockchains did a good job, but there are still a few common pitfalls that we see. Um, for example, we have decentralized apps.

You know, these are really strong primitives. They're backed by censorship resistant smart contracts as a database, but typically these DAPs are hosted on frontends that are um central. They're they're hosted on centralized servers. So if the front end goes down, essentially unless like you're very techsavvy, you basically lose access to the website. Um and we're not we should be targeting people without CS degrees.

So that is most users in most case. Um, and then NFTs, another really strong primitive that represents ownership on chain, but the metadata, especially images of the the NFTs that we have, they're hosted on centralized servers. So, you could wake up one day after having spent a lot of money on your NFT and the image is gone. So, I mean it you may have a copy of it, but good luck trying to sell it if you know the image is broken. And this is also an issue in web 2 as well.

like we all still like use the web two world the web as well and we have to trust these big data companies to act responsibly with our data we don't really own much um and I mean this is why we have blockchains to begin with we don't have we don't want to trust big companies um because they show that they can't be trusted so this is why we need decentralized data so like I said miss labs we built sui Um and our mission at miss labs is to decentralize the entire stack of the internet. So we see you know and other blockchains like as a global coordination layer. So computation coordination of all of these assets on chain um it's done that's that's really good but the decentralized web needs a global data layer and this is where Walworth comes in. We need a platform that is decentralized of course we need it to be reliable. We need it to be um cheap and we need it to be programmable.

And this is where Walrus comes in. So um I'm just going to give an overview of like all the different aspects of Walrus. So um on other decentralized networks, most of the data is replicated on every single node. Um but at Walrus, we do things a little differently. thanks to the the core walrus team that they created a new encoding algorithm that allows us to encode the data only a constant replication factor of four to 5x.

So this means that when you upload your data let's say a gigabyte we store four or to five gigabytes worth and then distribute that throughout the network. Um, so this makes us very cost- effective compared to other networks that use full replication or closer to it. And this also helps us with scalability because every node um doesn't have the same amount of data or same same data. When you add a node, you're actually increasing the performance. But on a network where every node has the same data, you're not really increasing um the performance there.

maybe the read data. But on Walrus, you can incre increase the read and write throughput by adding nodes. So even though we are we have a very low constant replication factor, we're still very reliable. The you'd still be able to write data to the network if a third of your network is down of the the Walrus network is down. And even if twothirds of all the nodes in Walrus go down, you'd still be able to retrieve your data.

So we, you know, we're we're trying to make sure that it's both cost-ffective but reliable because if we want people to trust uh the the reliability of the network and store valuable information, then they need to have some sort of guarantee that it's going to stay there. Um, also importantly, Walrus is programmable. Walrus, you may have heard Walrus uses SUI, the SUI blockchain as the coordination layer. So all of the metadata, the payments and things like this are managed on the suite blockchain. Um this allows us first of all to focus fully on the data part of it.

That's like how the core team can focus on the making it cheap and resilient. Um but this more importantly it also means that for the first time we have storage availability proofs as well as storage capacity as tokenized assets on chain. So you can imagine creating things like smart contract based blob storage um storage marketplaces, rental protocols or like integrations into DeFi and stuff like that. And most importantly out of all this, it's decentralized. That's the whole point.

Um so there's a wall token which is which powers the payments, the governance and all these things. Um and this is all lives on the sweet blockchain. So yeah, so this is exactly what we need and why I think Walrus is the best fit because we're cost- effective but reliable. You know, you should you can rely on us and we're also programmable. So we provide a lot of composability and a lot of opportunity for smart contract web 2 and web 3 developers to use this.

And it's decentralized. So um you don't need to trust a single entity or company for this. Um, and just a note as you guys are building and considering building, everything on Walrus is public by default. So when you upload an image or whatever, that's going to be public. So either you have to encrypt this data yourself beforehand.

We we didn't include any native encryption because we wanted it to be free form. We wanted layers above to provide the service. So we actually have another product called Seal which combines the encryption on Walrus with access control via smart contract. So you can have um you could create like a YouTube subscriptionbased service or a Netflix subscription where you upload your content to walrus encrypted with seal and you can create an allow list or subscription based based off payments on by crypto or even like an NFT. So like an NFT collection could post dedicated content for their users and only the holders would be able to view this and maybe even for like um a certain amount of time.

I'm not going to go into all the use cases we have but I wanted to showcase one that we have which I think is a good example of accessing the programmability of Walrus and for those of you building it might be a good tool for you as well. So Walrus sites is a platform that allows you to host frontends on Walrus. So this directly addresses the issue that I talked about earlier with DAPs being hosted on centralized servers. This um allows you to fully decentralize your app. So the smart contracts the backend per se is hosted on a chain and now for the you can also host the front end on a chain.

So this when your site's hosted on Walrus sites, it can't be taken down by any central agency. More than twothirds of the network would have to go down for this site to go down. Um yeah, so this just to go goes to show like how what are the opportunities to build and stuff like that. Okay, so I just wanted to give you guys an overview so you're familiar with what Walrus is, like why we built it and um why you guys should use it. So, more importantly, like at the hackathon, we have a track.

There's $10,000 in prizes total. And I want to explain the track and explain you what we're looking for and how you guys can get started. So, our track is split into two halves or actually not halves per se, but um so we have the Walrus developer tooling side. Um even though Walrus uses SUI for coordination, it's still chain agnostic. Anyone can write data and read data from Walrus regardless of what chain.

You don't even have to be on a chain. So we are looking for more tools to make the experience for nonswuite developers more seamless. For examples, in the past we've seen people build like bridging SDKs that allow people to like pay with Ethereum as the native token rather than for example. Um, this link is also on the Let me see if it loads here. This link is also on our sponsors page.

So, this just has some ideas that for you guys to look to. Um, these are specifically for the tooling track. And if you notice that some of these mention like an RFP program, this is essentially like a grant program. So, some of these we're actually looking for beyond the hackathon. So this could be a good opportunity to provide an MVP as your submission and that could help you guys um get this get future support as well.

So these are just some ideas. The crosschain support is probably the biggest one as you guys are EVM builders. Um ENS powered Walrus sites. Walrus sites allows you to use your SU name service domain for a custom domain of any site you deploy on Walrus. There's no reason why someone couldn't build an ENS powered one as well.

We already have the code that lets you host your own Walrus sites portal. You just have to add the functionality to map it to an ENS name. Yeah. And then on the other side of things, we have the best app with Walrus. So if you're looking to maybe combine this sponsorship with a few others, then the best app is probably better for you guys.

this. We're just looking for any sort of app that uses Walrus for decentralized storage. Um, when we're looking at when we're judging these apps, we're mainly looking at the app idea, the quality of your development, and things like that. Um, and I also have some app ideas. I'll just share a few.

Um, so like a decentralized link tree. Link trees are really popular, but it'd be really cool if your link tree was actually a dedicated Walrus site that you owned, not some site that Linkree could take down at any moment. Um, oh, and this is yeah, like Walrus Internet Archive. I'm sure you guys are familiar, but you could build one that's better, more decentralized, like actually decentralized decentralized Wikipedia. You could actually build it where the content's gated and you have to pay to get in and then you can actually pay the contributors of this content.

And I want to specify that I want to note that in the resources here I named some I listed some resources that are specifically chain agnostic. These are like so when you're building an app we don't want you to have to get bugged down by like using sui or walrus or gas token. So we have Typescript SDKs um an S3 compatible API as well as some other SDKs managed by the community that you guys can use that don't and you don't need like SUI to get started. Yeah. So this brings me to my next topic.

Oh, and just so you know, so the 10,000 is split into 2,000 slots. $2,000 slots. Three of those slots are for developer tooling and two are for best app with Walrus. Okay, so this is how you can get started with building on Walrus. Um, it's similar to how our tracks are split up.

There's two ways. You can integrate with Walrus and um, use technology from SUI or you can use more chain agnostic aspects which I'll talk about later. But I just want to briefly go over we have TypeScript SDK that you can use on both the server or client. Um, these clients can be used to upload and read data from Walrus and you could like have your um, connect it with like a wallet and have the wallet user pay for it. We have the wallet CLI or the Walrus CLI which you can read and write data from as well as the Walrus sites product has its own dedicated site builder CLI.

This is the CLI you would use to upload and update frontends. This the same repo that you build the CLI from is the one that holds the portal hosting um content as well if you wanted to build that ENS powered Walrus site. But more importantly, this is the these are the chain agnostic methods for building with Walrus. The very the easiest way is with the HTTP API. You can upload any data since this is on test net.

It's free. You can upload any data on Walrus with a simple put request. Um and this is can be anything like text, videos, images like whatever it's it you can upload anything that's arbitrary data form format. Yeah. So that's really easy.

And then we also have some partners who um built SDKs that abstract away the suite aspect as well. Um, Tusky which is like a decentralized Google Drive on Sui has a TypeScript SDK that also manages the endto-end encryption, file management, access control and you don't need any suite or walrus to get started. Not any anything on this page. You don't need any tokens for that matter. Uh, we also have the S3 compatible Walrit storage API by maintained by another partner.

So if you're familiar with S3, you can use that. And then we also have some community SDKs. Um, this is actually pretty cool. Like I mentioned like we have an iOS SDK. This was actually made by a builder at ETH global can.

So they were building in the tooling track and we really liked it. So we wanted to support them afterwards and we got it um and we're you know helping by sharing it with the community. So this is also good inspiration for you guys like if you wanted to help the community as well. This QR code points to a GitHub repo called Awesome Walrus, which has all of the third-party APIs that we offer. Um, this is also important if you're building a to building a tool.

You could use this to check like what already exists and what isn't there and what could be there. Um, yeah. Um, yeah. So, I wanted to stop talking for a bit and just have you guys try it out. So, you can take out your phones.

You don't need a wallet or anything. Go ahead and scan this QR code. And it will take you to a simple web app that we have where you can test uploading to Walrus. I'll give you guys a moment to do that. Keep in mind, like I said, Walrus is public, so don't upload anything sketch.

But yeah, you can go ahead and try that. Um, yeah. Yeah. And then once you have it uploaded, it'll give you like the link to see it on like the Sweet Explorer or the Walrus Explorer. There's a few options there.

Okay, I'm move on while you guys do that. All right. And before I end the talk, I just want to give you guys a few final reminders. Um even though like you you're going to hear like sui every once in a while but remember that walrus is chain agnostic especially in the scope of this hackathon. So if you're building with hackathon building a project with walrus and you're stuck trying to install some sui cli or trying to get some gas tokens you're probably down the wrong path.

So if you you can always double check with us like we'll make sure that you're using the right tools for your use case cuz especially if you're building in the app category you typ you won't need any suite or tokens um at all. Um another tip is awesome walrus. This is a really good list of the third party tools that we have available for you guys to use. These are the SDKs that don't require any suite integration. So these are the ones to use as well.

and also for inspiration. And I just wanted to also remind you guys that we're here to help. Like um my team and I will be at the booth like most of the time, maybe not at like 2:00 a.m., but any other time besides then unless you really need it.

Um unless there's a line, we'll stay. But so yeah, don't don't hesitate to come ask us. Um you can come up with ideas and see how what we think of them. Or you could ask for help, any developer help. Um or just check in with your progress.

We'd love to stay aware of stay up to date with how you're doing. So yeah. Um do any questions?

Yeah.

Sure. Yeah. Yeah,

sure.

Mhm.

Yeah.

Yeah. Yeah. So, like if if especially if like you don't care about the web 3 aspect, then we can't compete with AWS. like we're definitely an order and all like decentralized networks are like this an order of magnitude more expensive than uh like AWS. It's Walrus is really good if you care about like censorship resistance, data sovereignty like if you really like uh like a workout app or like a health app like I personally would care if like who owns that data and typically like I don't get the option to export my data on any of these apps but if like you can imagine if it's powered by a decentralized storage network you have the option like you're the only one that has access to this data.

If someone wants to use it they can pay you for it but besides that they don't have access. So, and at any point I can export it and move it to some other app if I find an app that has better features. So, I I like I know of people who I've switched like health apps and they just lose years of data because they can't control it. But there's definitely trade-offs for sure.

Yeah. Yeah. So, and I think blockchains really helped with the coordination part, but I think like stuff like Walrus can help on the data side. Yeah.

Yeah. Only for very special tooling, most likely tooling stuff, you'll need tokens, but otherwise it's free. This is test net and we have all the abstracted away stuff. So, yeah. Yeah.

Okay. Yeah. Um, if you're a user accessing a wall site, um, no. So, when you deploy a wall site, you pay for however long you want that storage to be available for.

Yeah. So, yeah. So the only thing you pay for is the storage and then there's a layer of like a network of people. It's a decentralized. It's they're called the portals.

The portals are the ones that take the static front-end code and render it on the browser. So like we host the portal um and like other people can do it. So if you have like special use cases or like a very high performant um high performant or high like available website like you might need to use like a more performant portal or just host your own portal. Um but essentially the data is always going to be on Walrus and once the ecosystem is more mature and we get more portals out there like you would have to run out of or all the portals would have to go down. So there's not a single point of failure at that point.

But I can I can we can talk. Yeah.

Yeah. Uh we'll Yeah, let's talk after. Yeah. Do you have

a question?

Uhhuh.

That's a good one. I honestly So we launched mainet in March, so we're still pretty young. Honestly, like I think my favorite would be the one that hasn't built yet. So, I think like the health app is another example. I like because I personally would really like a health app where I actually own my data.

Um, and like be able to export it so I can use like other AI features that I want to build around it, but like right now I can't do that. So, it's a it's a bummer. Like a decentralized publishing platform, stuff like that. Like, yeah. So, I I I think there's still a lot of opportunity.

So, a lot of the cool things that I want to want to be built aren't built yet. Yeah. All right. Yeah.

No. Yeah. So, this is a little different if you're familiar with IPFS or something like that. Um, you pay for how long you want it to be stored. So, um, basically we want to be realistic and like I don't know the forever thing.

It's hard to buy sometimes, but so like an epoch on mainet is two weeks. So, you have two week increments. You pay for how long you want it to be and at any point before it expires, you can renew it. Um, but at at any point after it expires, it's not guaranteed to be on the network anymore. We're working on services like backup services and stuff like that because it is a new paradigm because people are used to paying a subscription and it's there and then I as long as I pay it's there and even like when people miss payments the companies still keep the data there so you can like catch up.

So, we're working on tools to improve the experience there. But

you you can delete it if that's what you're asking. as a storage node. Um,

yeah. So, the store it's kind of like a blockchain like the the nodes can decide as a network if there's content like illegal or something that's they don't want to show. Yeah. Yeah. Mhm.

Is it working? Yeah. Oh, okay. Well, I can take a look after and see. It was working earlier, but yeah.

Oh, yeah. I I think there's a there's a limit on the site. It says 10 megabytes, but the publisher Yeah. We'll we can talk after um because it should be working. It might be a size image like on test net there is size limits because it's all free.

So

yeah and if you you can bypass it on test net too. There's ways I can show you how to do that. Yeah. Yeah. So and yeah the max file size that we have per upload is 13 gigabytes like in real life not on test net.

So yeah. All right. All right. Well, thank you guys all so much. I really appreciate the questions and I hope to talk to you more about this.

Yeah. Good luck, guys. [Music]

Automatic transcript — names and jargon may be misspelled.