Jason Goldberg I Airstack Product Demo | Pragma Lisbon 2023

foreign [Music] thank you everyone first of all um everything we're going to do today is live so if it doesn't work so be it we're going live everything's live on the internet um I'm going to show you some things today you've never seen before everyone here has probably heard of chat gbt everyone's tried chat EBT okay so today I'm going to show you the first searching of blockchains using AI if you've ever gone to church EBT and typed in say show me all ens names of nouns Dow holders the answer is I have no idea with airstack we have built the ability to do natural language searches on blockchain in a way that enables developers to easily search cross chain cross dap on chain and off-chain data and combine it into a single query in response you all ready all right so um here you can see the airstack website uh the way we position airstack is data apis for Builders a lot of people when they think about web3 data they think about analytics we think about web3 data from a builder perspective I'm building an application and how do I integrate web3 data blockchain data data about what's happening across the weather ecosystem into my applications background on myself and our team we've been building on ethereum since 2016. in fact I remember back in 2018 meeting Patricio who we just heard from at events in Waterloo where he was handing out people physical co-ops the first nfts one of the things that we observed over the last six years while building on ethereum I guess almost seven years now is that the technology has gotten progressively better we used to go from having to you know run our own nodes with with the tools like infuro to then running sub graphs with the graph to then using apis from the makers let's say Alchemy and quick node and Morales but one thing that we heard in conversations with developers the last couple years is that while things have gotten better we're still a long way to go before developers can literally just ping an API and plug and play data from across the ecosystem we heard from developers that if they want to integrate data that's outside of their own application they had to run all sorts of fire hose nodes ping all sorts of whether it's subgraphs or other Technologies build complex back-end databases analyze the data and then serve it up to their customers and with airstack what our design is to enable anyone to be able to just use the airstack apis and easily integrate any data into their application without requiring any back-end system on your side so literally our vision is I'm going to show you how real it is today that you could be building an application let's say it's an application to be able to send marketing messages or airdrops to customers and you could say I want to only I want to First do analysis of everyone who has been to pragma and who maybe also went to Devcon and you'd get that by analyzing say their co-ops or the nft tickets and then also maybe you played a game on the sandbox or owns a board ape I want to just pinpoint those users and send them this message prior to airstack almost impossible to do you'd have to do to Ping a whole bunch of apis download a lot of data cross-reference it but now I'm going to show you is this is now possible so if I go out to airstack and first I just let you know that everything I'm showing you here today is all live on the Internet it's at app.airstock.xyz and you can use it for free we're purposely not charging bear stack right now we just want to gain use cases get lots of developers using it we launched at e-tokyo a couple weeks ago we have now over 300 development teams building on airstack and the number one thing we heard from them is that they were able to build more faster better with airstack than they were before because they didn't have to spend days or weeks just Gathering data instead they could focus on their applications so when you go to airstack the first thing you're going to see is there's a community built into the product so whenever someone runs a query on airstack so for instance get me the recent transfers of die tokens you can publish it at the community so other people can examine that query they can use it they can Fork it themselves we are live on the internet and the internet's not working okay let's try again here here we go all right so here for instance is that query all right it's a little slow see we can do here we go um here for instance that query that someone did on uh tracing transfers and die tokens and I can run this query and get a response here you see the query and the response uh in a few milliseconds now what I want to show off today is the AI capabilities because this is what really makes this special this is what's going to open up web3 developer tools to not just the 10 000 or so ethereum developers today but to anybody who has a question about a data that's on blockchain so the first thing is uh on airstack you can query today ethereum polygon you can query across all the all the major web3 socials so you can query across ens forecaster lens nfts and you can also query across nft marketplaces like openc and wearable and blur so let's show a couple queries so the first thing I'm going to do is I'm going to ask the airstack AI to show me all tokens held is everybody able to read this okay held by the metallic .eath um and his web3 socials um I was saying also let's just do all ens held by vitality and three socials so what's Happening Here is our AI engine which has been trained on our apis is now searching across the the apis on airstack which have everything from domains to marketplaces to tokens to Wallace to balance transfers uh to and it's as I said across uh ethereum polygon ens lens forecaster and more so here you see the AI build this query from right for me it says the identities of italic.

eath so it's able to translate his ens name and it's looking on blockchain ethereum and it's looking for his various domains as well as his socials so I'll run this and we'll see okay so metallic owns these ens names these are his actual ens names we also see that his primary domain is setsu satoshiinart.eth probably didn't know that but that's how he said his primary ens you can see that his forecaster username is V buterin and his lens name is vitalik.lens so the AI was able to interpret this and return this back in a single query and if you think about what this query is combining it is combining data across chain crossdap so it's combining data that's on ethereum uh it's created by the ens as adapt it's combining forecaster as adapt that's actually on Gorly and it's combining dads on lens and a single query and response response in in milliseconds now in this case I started with the ens address let's do the reverse so let's start with the lens let's say show me all nfts held by stani.lens um and we'll generate that query so here I'm starting with the lens address and the airstack AI knows that this lens address is associated with an ethereum address is now also looking for all standings nfts across ethereum and polygon so it's building this query for you now one thing we found is that as I said there's about 300 developers that are currently using airstack after being live for about four weeks and we're finding that even amongst the savviest developers who are really familiar with writing graphql queries that over 80 percent of all developers are starting with the AI and why are they doing that because it's easiest way to express your problem statement or your use case is to write a question in natural language and then you can also then use the the apis to add more variables but the AI is pretty much 99 plus accurate on anything that's in the air stack apis right now so here I'm going to run this query and I see here's Danny's token balances yes steady.lens he also has these other lens names you might not have known that the lens official.

lens is registered understand these ethereum address as well I can see nfts that he's holding and not just the nfts I also can return back from airstack resized images of the nfts that we're powering through our CDN so if you think about again this query is across ethereum it's across polygon it's Gathering data from ipfs it's combining all this in a single queried response I'll do one more on web3 identity let's do for forecaster user uh gwr Dan Romero dwr is the founder forecaster I'm going to say show his lens and ens now I'm doing this with a single input but you know you as a developer you can put in variables you can put in thousands of addresses and the use case here for instance we have let's say owners of dows who come to me and said like tally uh Dennis and Italy said well if I want to see the ens names of everyone who's used this Dow and then also show what co-ops they've earned they can just do that in a single query on airstack so here I have for this forecaster user dwr I have here's this forecaster profile dwr his lens name remember.lens his ens addresses all in one querying response so if you think about developers who've been building over the last few months or if you they'll say a few like let's a year on the web 3 socials whether it's lens or forecaster until today they had to pick a lane essentially you build something on farcaster or you build something on lens or maybe built some of that ens now with airstack you could build interoperable applications that's one of the things we're really excited about for the hackathon this weekend is to help developers build applications that cut across the web 3 socials um so here you see an example of you know by just pinging the air stack apis the AI can help you and you can actually get all the social identities of someone based on their address now we can do more with airstack than just talked about uh web 3 identities we can also go to owners of nmt collections we're owners of Dallas or members of Dallas let's do a dow for instance so I'm going to say show me uh holders of Orange so one thing we've done is we've normalized the for over about 5 000 tokens we've normalized seen this at mentioned name of any major nft collection or token so I can say show me the holders of orange Dao and their ens and lens names and I can just generate this as a query here and the AI is going to build this query for me that's looking up it knows the address of the the orange Dow orange down for those who don't know it's a YC y combinator uh alumni investment owl and now I'm gathering across uh and and by the way that that token is on polygon and now I'm getting the addresses uh and the ownership of those these of the owners there and here I have uh token balances and whether this person here this uh owner of orange Dow token is buyer institute.lens here's another one that is forecast user Fahim all back in the same response show you one other uh query you can for instance query across chain um show all owners uh let's go Trump let's go Trump digital collectibles and their ens names not sure everyone have any hits here but we'll see um and so um I do believe there's a lot of ens Quarters on Trump digital Collectibles we'll see so here I'm querying again from polygon uh back to ethereum and ens and getting all the owners of a nft collection uh and their ens names and so uh we'll see what comes a bit and I'll show you one last one before we go actually so here we have owners of the Trump digital Collectibles um and not many ens names there but lesson said and let's do a resized uh nft images and here we can get back the actual owners of the token plus images of the nfts in four different sizes so application developers no longer need to say get the addresses of a contract holders and then look up the uh nft images on ibfs download those images resize them build their own CDN um you know manage all this through the AWS and S3 instead they can just ping airstax apis and just literally integrate it directly into their application so here I'm going to say all right here's owners of the Trump digital Collectibles and then for each one here is images let's see Trump come on let's do it all right with the internet connection is not loaded here but you can see there are images here usually it would show a preview here right now so I want to show you one last thing for this presentation is we have been there you go there's Trump all right so you can see uh there's our almost favorite ex-president um just kidding uh so last thing I want to show you is um our team has been building an app called Jam uh we started off building a jam on Far Caster but we'll be extending to other networks um like lens um and and others um and we've done it as a Showcase of air stack apis and the capabilities and tomorrow we're launching an integration of everything I just showed you here I'm going to show you a preview on my development server of that app so you can see just what the possibilities are with what you can do with web3 social when you start integrating all these components into an application so let's go here and share my iPhone all right so this is Jim uh Jam is an application built on farcaster we have within Jam there's various speeds there's this is actually feeds of tasks that are happening on farcaster in real time this is my for you feed which is customized specifically around what I'm interested in I can also see a live feed of every cast on Far Caster as it happens you can see there's one from Lisbon just now live that was a good time if we ever did that thank you um and we also have things like AI powered search built into the app what I want to show you today is actually the development staging version of the app so a couple things you'll notice um if you look at the user profiles there's a instead of the familiar let's say with the Twitter blue verified badge instead of using the verified Twitter badge or like the verification check mark what we're using is if someone has a primary ens name set we're verifying with ens and that's who that person is because I believe that ens is people's number one identity when it comes to web three and so you can see this user plugs in I'm going to click in on his profile and you see that his ens is pugston.e so I can verify that that actually is him because I see the ens badge and I know that is not eat so if someone was trying to fake that their vitalik for instance they can't because only vitalik is metallic.eath right next what you see here in terms of the profile is the profile is a collection of where that person collects because I believe that or we believe that whereas in web 2 let's say you are what you should post and we have three you are what you collect you're the events you've been to the co-ops you've collected the nfts collected the digital Collectibles that that represent you and that people are going to start using nfts in more ways let's say to represent the projects they work on the companies they're part of um and to start to build these personas around the nfts that represent who you are on chain so here you say for pugs to not eat you have his his forecaster nft is ens you see he's a has a collection called far cans sound.

xyz blobs and interface and just to see when you click in here the user is able to see everyone who has that collection so everyone is a member of that ft um and also decide which of them they want to add to their feeds as well so as we start to build new social feeds this building do social feeds around let's say everyone who works at gnosis or everyone who's been to pragma or everyone who collects a far can and then also the next step is then being able to send messages to and from those people through both things like forecaster and lens so I hope I showed you today some of the power of what we're building an air stack it's free to use give it a try have fun with the AI the best thing about the AI is the more people use it the smarter it gets because every time it makes a mistake we feed it right back to the AI and it learns more so thank you everyone appreciate it