--- title: Farcaster Batches Day 1 - builder showcase (GM Farcaster) show: Farcaster Batches episode: 1 guests: - Chris Dolinsky (Vinny App) - Kenny (POIDH) - Nikki Sapp (Juke) - Jonathan Colton (Founder Check / Fotocaster) - Dr. Deeks (email remittance pro, Crafters, Ghostwriter) host: Adriene and Nounish Prof (GM Farcaster) co_host: Adrian (diviflyy) - co-founder, Empire Builder organizer: JubJub date: 2026-06-02T00:00:00.000Z format: audio-space-showcase language: en track: builder topics: - farcaster-batches - farcaster-ecosystem - vibe-coding - mini-apps - on-chain-bounties - audio-spaces - founder-validation - artist-sovereignty - tokenomics - clanker - empire-builder entities: orgs: - GM Farcaster - The ZAO - Farcaster people: - Adriene (GM Farcaster) - Nounish Prof (GM Farcaster) - Adrian (diviflyy) - JubJub - Chris Dolinsky - Kenny - Nikki Sapp - Jonathan Colton - Dr. Deeks - Erik - Zaal - yerbearserker projects: - Vinny App - POIDH - Juke - Founder Check - Fotocaster - Crafters - Ghostwriter - email remittance pro - Empire Builder - Clanker links: - label: POIDH url: https://poidh.xyz - label: GM Farcaster url: https://www.gmfarcaster.com source: capture: live Farcaster Batches Day 1 Space recording (audio), auto-transcript cleaned (proper nouns corrected, filler trimmed, dropout hallucinations removed) summary: | Day 1 of Farcaster Batches, a week-long builder showcase across the Farcaster ecosystem organized by JubJub and hosted by the GM Farcaster team (Adriene and Nounish Prof), with Adrian (diviflyy) of Empire Builder co-hosting. Five builders present and field questions: Chris Dolinsky on Vinny App (natural-language app builder on a Web3 stack, running a $1,000 Vinny-a-thon prize pool, extra $250 added by Justin/AHN.eth); Kenny on POIDH (a fully on-chain social bounty platform now live on Ethereum mainnet, with a 20-bounty $25 boost promo for the week and the $30,000 Farcaster-crowdfunded kickflip world record as its signature win); Nikki Sapp on Juke (an open-source, audio-native Farcaster client that helped push Farcaster toward native Spaces, now shipping voice notes and an audio SDK that Zaal helped test); Jonathan Colton on Founder Check (an agent-driven GTM/validation mini app from the "Distribution Is Hard, Don't F It Up" book) and Fotocaster (a maximum artist-sovereignty photo mint built with Erik, ~6,000 collects in 30 days); and Dr. Deeks on a slate of mini apps - email remittance pro (send crypto by email, gift-card payout, 1.5% fee), Crafters (an AI word-combination game with Megamind NFTs), and Ghostwriter (a Madlib-style community storytelling game). The throughline every speaker and Adrian return to: build the thing you want to see in the world. action_items: - "ZABAL Gamez builders: Vinny-a-thon runs Jun 1-5, judging Jun 5 - build a Vinny app (optionally + Empire Builder) for up to $1,000 USDC" - "POIDH: first 20 Farcaster-tied bounties this week get a $25 boost + promotion; create on Ethereum mainnet (min 0.001 ETH)" - "Juke: open-source repo + audio SDK open to contributors; integrate native Farcaster Spaces into a site or app" - "Founder Check: DM Jonathan to run an idea through a live workshop session (Tue)" - "Watch the rest of Batches week: Tue Founder Check workshop in the GM Farcaster studio; Thu Cashless Man + Royal Aid on AI automation and self-hosted infrastructure (Farcaster Spaces)" status: cleaned --- ## Key points - **Farcaster Batches** is a week-long, ecosystem-wide builder showcase organized by JubJub (asleep in Australia during Day 1), hosted by the **GM Farcaster** team. Day 1 features five builders; two live workshops follow later in the week. - **Vinny App** (Chris Dolinsky): prompt a fully functional app into existence on a Web3 stack in minutes, any language. Each app gets its own EVM wallet; optional Clanker token whose trading fees fund that app's AI build costs and roll 10% up into the umbrella Vinny token. Vinny-a-thon: $1,000 USDC prize pool (extra $250 from Justin / AHN.eth), build Jun 1-5, judging Jun 5. Stack composes with Empire Builder (yerbearserker / Diviflyy). - **POIDH** (Kenny): a fully on-chain social bounty platform - crowdfund any outcome, verify it trustlessly. Each bounty is an "ephemeral DAO" with low attack surface (one specific yes/no outcome). Now live on Ethereum mainnet. Signature win: a $30,000 Farcaster-crowdfunded kickflip world record. Week promo: first 20 Farcaster-tied bounties get a $25 boost. - **Juke** (Nikki Sapp): an open-source, mobile-native, audio-first Farcaster client. Built in a week to push for native audio; Farcaster (Rish) later shipped native Spaces. Juke keeps going for the open-source baseline client plus features Farcaster lacks - voice notes, recording, and a new audio SDK (Zaal helped test it end to end). Deliberately no token. - **Founder Check** (Jonathan Colton): an agent-driven GTM validation mini app grown out of the "Distribution Is Hard, Don't F It Up" book - "Duolingo for builders." A live session with Zaal jumped his score 3.3 points out of 10. - **Fotocaster** (Jonathan Colton + Erik): a photo mint built for maximum artist sovereignty - artists keep ~100% of proceeds, platform takes a penny per sale. ~6,000 collects in 30 days. - **Dr. Deeks**: email remittance pro (send crypto by email, recipient claims to a token or gift card across Monad / Base / Celo, 1.5% flat fee, built for the Synthesis hackathon), Crafters (an AI word-combination game with first-creator Megamind NFTs and agent-driven features), and Ghostwriter (a Madlib-style community storytelling game with hidden-metadata NFTs). - The throughline every speaker and Adrian keep landing on: **build the thing you want to see in the world.** ## Transcript > Cleaned from an auto-transcript with no speaker diarization; speaker labels are inferred from > context. The Space had two audio dropouts (an X / Spaces glitch mid-session and Dr. Deeks' > phone overheating near the end); the dead-air sections, where the transcriber looped, have > been removed. Some names are best-effort and listed in the glossary's "Confirm before use." **Adriene:** GM, GM everyone. Give a recast, let people know we are here. We are recording in our studio as well as in the Space. Adrian, I'm making you a co-host so you can bring people up for questions. We're waiting on a couple more folks - Nikki, Jonathan, Dr. Deeks. Let me do a quick mic check with everyone on stage. Jonathan, Chris, Nikki, Kenny - good. And I'm bringing up the Farcaster intern too; they don't have to speak, but it helps visibility. All right, folks. We are here for Farcaster Batches - I almost said Base Batches - Day 1. This is the brainchild of JubJub over on Farcaster. JubJub is in Australia, so he's actually asleep, but he did all of the organization for this. He wanted a week-long focus on the builders around the Farcaster ecosystem - pulled together all the builders and two workshops happening this week. Tomorrow there's one in our GM Farcaster StreamYard studio; Thursday we have two builders, Cashless Man and Royal Aid, doing workshops on setting up AI automation and building on your own infrastructure, both on Farcaster Spaces. A lot going on this week. Today we have five amazing builders: Chris Dolinsky with Vinny App, Kenny with POIDH, Nikki Sapp with Juke, Jonathan Colton with Founder Check, and Dr. Deeks on several projects. Each builder gives us a breakdown - what your project is, what you've been building, how it connects with the Farcaster ecosystem - then we'll take questions. Chris Dolinsky from Vinny App, I'm going to start with you. Welcome. **Chris Dolinsky:** Great to be here - always wonderful to see so many people on the internet jump into one Space. Vinny App is a platform that lets anyone in the world, developer or not, use natural language to build a fully functional application in a few minutes. It's set up on the Web3 stack, so it makes use of brand-new primitives that have never existed historically. You prompt your way into having an app. As part of Farcaster Batches we have a $1,000 prize pool - Justin, AHN.eth, contributed an extra $250. Anyone who builds during the next five days is eligible for up to $1,000 USDC for any app built using Vinny - either a pure Vinny stack, or with Empire Builder (yerbearserker or Diviflyy) or Justin's project added in. You can be anywhere in the world, walking on a treadmill or getting a coffee, and start prompting your build. There's a full leaderboard so everyone can see the apps being built. We track unique users per app, and each app in the Vinny-a-thon launches with a token, so the market can get involved and the market cap is trackable. Building runs the first through the fourth; on the fifth we judge live and award the prizes. The real goal is to take the narrative that everyone should be building and actually open it up to the 99.5% of people who historically either can't code or, more often, just don't have time to string all the pieces together. It works in any language - Korean, Japanese, Mandarin, Portuguese - we have apps built from Turkey and India. Half the time I have to run the prompt through Google Translate to see what they built. While you build, you have full support from our team - gamification, tokenomics, features, bugs. And 10% of the trading fees of every launched token roll into the main Vinny token, so there's a canonical umbrella token binding the platform together. It's a land-and-expand strategy: start with a little nugget - a card game, a health tracker, an expense tracker - and grow it exactly how you see fit, then share it as a mini app on Farcaster or out on the wider web. We have World support (around 20 million wallets), Celo MiniPay support (another 15 million), Telegram support coming, and direct web use. Really looking forward to seeing what people build. **Adriene:** That's awesome. I'm definitely going to try to build something. Adrian has built a couple things on mini apps - what tips would you have for people participating this week? **Adrian (diviflyy):** I've used Vinny twice since Chris demoed it and it's really awesome. Don't just use it - try it. If you get stuck, ask Chris or Nick for support, they'll help out. I found it super intuitive, really easy from idea to prototype. Chris, I have a question if you'll take one: part of the flow prompts people to launch a token before the app is built, and it's framed as fees from the token helping support the build. I like that model - have you gotten any pushback or confusion about what the token does? **Chris Dolinsky:** Totally. The tokenomics make use of these new primitives. There are a number of vibe-coding platforms, and people will say "oh, they did that on Claude Code or Codex." Good to know: with Vinny we sit on top of those - you take the bare bones of a Claude Code or a Codex and layer on a ton of skill sets and connections you'd otherwise have to provision yourself (API endpoints, wallets), and eventually you get to the idea of a token. A token isn't required to use Vinny as a standalone platform - it's only required for the Vinny-a-thon, because it helps with the leaderboard and tracking and lets the market participate in ranking the apps. What's nice: every app that gets created gets its own EVM wallet, which pays for the AI building the app, storage, image generation - basically everything. As the app's token trades on the open market, those fees accumulate into that wallet, so for some builds the trading fees actually helped pay for the development. But the real purpose is to let people explore how tokens can be used in novel ways. A chess app uses the token to buy animated boards; Nick's app Uchi uses the token as a tipping system. Because it's on Base it's light and cheap to transfer, and with Farcaster wallets and Clanker tokenomics it's exciting to see which direction builders take. **Adrian (diviflyy):** Love that. Great answer. **Adriene:** We're running a little long already - shocking, but that's okay. I'll point out that you and Nick came on GM Farcaster a couple weeks ago and built a Pac-Man game live on the show, so if anyone wants more instruction, check that out on GM Farcaster's YouTube. Any closing thought, Chris? **Chris Dolinsky:** It's so easy to get started - don't carry the weight of the world. Just type your prompt: "I want a Pac-Man game," and start there. Don't worry about all the functionality up front. The front door should feel like walking into a spa - calm, no pressure, just look around. You get a little momentum and a couple weeks later you have a fully functional app, like the 2048 Farcaster that Adrian made, where the whole feed was posting about it. **Adrian (diviflyy):** For a few hours a day. **Chris Dolinsky:** Hey, Andy Warhol said 15 minutes - don't expect more. **Adriene:** Thank you so much, Chris. If you have questions about Vinny App, reach out to Chris. We're going to move on to Kenny from POIDH - and Kenny, give me a better pronunciation. **Kenny:** Everybody's first question. Early on I said POIDH, but lately I just say POID. POID has been running a while - I've been building it in public on Farcaster for three years. POID is a social bounty platform: you can crowdfund any outcome, verify it happened, and do it all trustlessly, 100% on chain. There have been lots of task platforms in crypto, but our twist is that when you create a bounty you're creating an ephemeral DAO - as other people put money in, there's this aspect of people rallying to make something happen, and you actually see it on chain. For this week - and first off, thank you for the CEF grant last week, we're stoked - the promo that coincides with Batches is an activity grant: the first 20 people who create a bounty tied in some way to Farcaster or your project get an extra $25 plus promotion, just so people get a feel for the app. Our minimum is 0.001 ETH, so you need about $2 to start. POID is a broad design space - you can incentivize almost anything: usage bounties to drive users, competitions, clipping bounties (we did one for GM Farcaster two weeks ago), bounties to summarize your project, even IRL guerrilla marketing - stickers and posters. We did one for QR coin putting up stickers around the world, and one of my favorites was someone putting an "Ethereum for financial infrastructure" poster up on Wall Street, crowdfunded through about 10 people for $200. That's the spirit of POID. And we're live on Ethereum mainnet now - contracts tested and looking good - so we'd love to see bounties created on mainnet this week. **Adriene:** Love this. Where did the concept come from, and what have been the most interesting bounties? **Kenny:** Years of ideating. I'm a big governance and DAO nerd - when I first heard about crypto that's what caught my eye: how can you use this to create online communities? This came out of the pandemic era - I was thinking about ways to crowdfund for public goods and healthcare. The fundamental problem was permissionless crowdfunding where anyone can put in funds and anyone can claim - but if claims and voting are permissionless, people could vote to take the money themselves. So I narrowed it: a DAO for one specific thing has way less attack surface, and the people who put money in vote on whether that one specific thing happened. There are lots of governance failures when you vote on subjective things, but people on the internet are pretty good at agreeing on "did this one specific thing happen," especially when they have money on it. The big one, made possible by Farcaster, was the kickflip world record - we crowdfunded $30,000 basically within the Farcaster ecosystem, and Dave Bachinsky broke the world record for kickflips in one minute. That's the one that made it crystal clear we had something special. I just posted an article called "The Ephemeral DAO Machine" that runs through 10 of my favorite crowdfunded POID bounties - and one big failure, because no tooling is perfect. **Adriene:** Awesome - everyone go check out the bounties going up this week with the bonus. Adrian, any questions before we move on? **Adrian (diviflyy):** Nope, that was great. We love POID. **Adriene:** We love POID - and I'm bummed we know how to pronounce it now, it's more fun to guess. Kenny, kudos to you, you really are the epitome of building in public on Farcaster. **Kenny:** I'd say you all were the first ones, but I appreciate that. [Technical difficulties - the Space glitched on X and dropped several speakers; a few folks cut out and came back over the next minute.] **Adriene:** Okay, we're back. Nikki, give us the overview on Juke. **Nikki Sapp:** Appreciate you hosting - sorry for the technical difficulties, that's Elon's fault, never happens on Farcaster Spaces. So Juke came out of necessity and desire. I casted about wanting native audio Spaces - I missed Tavern from back in the day. When I first joined Farcaster, Tavern was huge; there was Farcaster Audio and Farhouse and all these attempts to bring audio to Farcaster, but they each had pitfalls, especially in the mini-app context - if you soft-closed the app or your screen went dark, you'd lose the audio. So I felt a native experience was necessary, there was demand for it, and I just built it in about a week and published it as an open-source, mobile-native, audio-focused Farcaster client. I did it partly as a poke to Rish and the Farcaster team to get real native audio Spaces. I won't take credit - several others were doing audio-focused things too, and it all converged. Rish reached out and asked, "would it kill you if we add it natively?" and I said no, it'd make me super happy. So Farcaster has Spaces natively now. Why keep working on Juke? First, it's important to have an open-source, mobile-native baseline Farcaster client - point your LLM at my repo and you can build your own client simply, with the audio features. Second, Juke has features Farcaster doesn't - voice notes you can post directly to the feed (they render natively on Juke, as a mini app elsewhere), Space recording, and a new SDK - which Zaal, who was in the audience recently, helped test end to end - so you can integrate native Farcaster audio Spaces directly into your own site or app. Cassie has unified audio across quorum and the main Farcaster app, so I'm working on syncing that for Juke too. My goal isn't to compete with Farcaster - this is a passion project. As Farcaster ships features I'll add them; as I come up with new stuff I'll add it. The app supports mini apps, snaps, and the usual protocol-native features like channels. The one thing it doesn't support right now is crypto - no wallet integration yet. That may change to support things like mini apps, but I'm not adding a wallet to launch a token. I have no plans for a Juke token - I have no desire to take on that level of smoke again unless it's very, very worth it. Honestly I use Juke a lot myself because it has a GIF keyboard and things I wanted in the main app - and so can you, just fork the repo. **Adriene:** That's incredible - the personification of "build what you want to see in the world." Are you looking for contributors? **Nikki Sapp:** Absolutely. Zaal has been super helpful - he helped test the SDK end to end. There's an open-source repo, and I'm happy to let people into the nightly builds. Ideas, feedback, build your own client on top of it - I'm flexible and open, so reach out. I'll just say: there's still a massive opportunity to build on Farcaster. Sentiment shifted with the change in ownership, but the primitives and tooling are massive, even if you're just a vibe coder, especially as we see the closed-loop centralization of other platforms. Don't give up on Farcaster - it's very viable and keeps improving. **Adriene:** Love it - and you've been one of the big supporters while giving honest feedback all along, the team appreciates that. We're going to move on to Jonathan Colton. Dr. Deeks, I see you in the audience, I sent you an invite - we'll chat with you after this. Jonathan, welcome to Farcaster Batches. **Jonathan Colton:** Founder Check was a labor of love. It started last year when I published V1 of "Distribution Is Hard, Don't F It Up" as a Google doc and shared it with the founder channel on Farcaster - a slightly embarrassing version, which I think is good; if you're not a little embarrassed you're probably doing it wrong. I released a V2 after talking to builders at Farcon base camp for months. Then earlier this year, when we all figured out we could do cool things with agents, I started doing GTM pain research on Reddit and Product Hunt, building a corpus and knowledge base, with a hypothesis: what if lean validation could be done by agents - could they discover signal faster than the founders who need it? I built a first version on Replit, then rebuilt it with the Farcaster building agent - way faster and a better version, with a built-in wallet and my social graph attached. People know me, they know I wrote those distribution versions, so I had some trust and credibility on the topic. Once I launched, I really just paid attention to the analytics - how well people performed in each pillar of Founder Check - and that data helped me create interventions where people performed poorly, which became main features. **Adriene:** Tomorrow you're doing a workshop - what's the goal, who should tap in? **Jonathan Colton:** Tomorrow I'll go into the user pain and inspiration, then run a live workshop: we select one idea and run it through Founder Check live, you see the score reveal, and we might do a quick revision to see the score improve. There's a whole "Duolingo for builders" side - builders come in, use Founder Check, improve their score. I did a session with Zaal last Friday and his score jumped 3.3 points out of 10, one of the top jumps I've measured. The outcome is that founders and builders get better at communicating an idea clearly enough that people can understand and replicate it. If folks want to work through it live tomorrow, send me or the hosts a DM and we'll pick one to do live - that's always the most revealing version. **Adriene:** Very cool. While we have you - talk about Fotocaster, it's been popping off since it launched about a month ago. **Jonathan Colton:** Fotocaster was a byproduct of Erik and me complaining, in private and a little in public, about minting your work and what sovereignty artists have. Everybody remembers Rodeo - we were giving up 50% of a sale and had no sovereignty over editions or one-of-ones. It was a one-shot model and it worked great; I minted 143 NFTs and had over a thousand collectors - but Erik and I always felt something was missing. With Rodeo gone, the human drives to create and to collect were absent on the timeline. Three weeks into building Founder Check I begged Erik: "it's time, we've got to build Fotocaster, and you have to use the building agent because it's amazing." Erik was the architect with direct access to the building agent, and we worked together every day for hours. There have been over 6,000 collects in the last 30 days. It's inexpensive - a joyful way to collect art you like, mostly photography - and the big idea was maximum artist sovereignty: artists keep 100% of their proceeds and the platform takes about a penny per sale, which we'll have to change to be self-sustaining, but we wanted to give artists maximum revenue. **Adriene:** Love that - there's been so much fun photography and art on the timeline because of it. Adrian, any questions before we move to Dr. Deeks? **Adrian (diviflyy):** No specific questions - I just appreciate the theme running through all of these speakers: build the thing you want to see. It's the most powerful thing to do. **Adriene:** Jonathan, we'll hear more from you tomorrow during the workshop - DM Jonathan or tag us if you're interested. Last but certainly not least, Dr. Deeks made it to the stage and has a number of projects - ghostwriter, email remittance pro, Crafters, and Journey Through Time. Where do you want to start, and how did you get building on Farcaster? **Dr. Deeks:** I'm not too big a fan of social media, honestly, but Farcaster wasn't just a social platform to me - it was a collective of individuals who want to grow and continuously evolve and help others grow with mini apps. Let's start with email remittance pro. It started for the Synthesis hackathon - a 10-day sprint that was supposed to be fully autonomous-agent-created, so I didn't write a single line of code. I had technical issues for the first seven days, so I only had three days, but I got a demo produced. The idea: no matter who it is across the globe, if you have their email you can send them crypto - the email is the identity layer. It's not required that you or the recipient verify identity, though that comes with its own risk, so there are two modes, business and personal. Personal: open the app, put your email in, connect your wallet, send funds to the escrow wallet, and the recipient clicks the claim link in their email. They choose how to receive it - the token sent, a different token on Monad, Base, or Celo, or a gift card - and any swap or bridging is done on the back end. For the recipient it's just click, select, instantaneous payout. The concept came from how businesses like Wells Fargo skim money off people sending funds to families abroad. Our model is a 1.5% flat fee taken when money is sent in; everything else is on the user. It's already proven on chain - I have transaction proofs and video showing it works. I'm currently redeploying it on Render and Cloudflare for the front end and back end. **Adriene:** That's super interesting - especially trading it in for a gift card for someone who may not have a wallet or want to deal with that. It also lets the sender skip buying a gift card or using traditional methods if they're crypto-native. Feels very early. Where do you want to go next? We've got three more. **Dr. Deeks:** We don't have to do all of them - my phone was overheating, it's been a chaotic Monday. Let's do Crafters. It's a word-combination game: everyone starts with a base set of words - water, earth, air, fire - and you have a draggable playground where you combine two words and AI combines them into a new word, documented in the database. The first time a word is ever created, you have the option to mint a Megamind NFT that says you're the creator of that word; after that, nobody else gets that NFT for it. I'm expanding it with agent-driven features - a Megamind can be poached in a "Megamind heist," there's rival coaching, you can rent out agents for brain power so they keep trying to make words on your behalf, daily challenges, on-chain check-ins, leaderboards, and a dynamic celestial weather system where the scenery changes based on crafts made or agent actions. **Adriene:** Is this a mini app, and how do you play it? **Dr. Deeks:** It's a mini app or web app - I have a beta as a mini app already, but I have a lot of projects I'm working on. **Adriene:** That's really cool. There are a lot of word games on Farcaster - shout out Word a Day - and they do pretty well because a lot of people dive in and play. Adding the agentic layer will definitely drive interest. I'm a little short on time, but one more opportunity - any other project you want to highlight before we wrap? **Dr. Deeks:** I'll touch on Ghostwriter real quick. It's a community-driven storytelling game, Madlib-style: you contribute a word into any active story to get a creator credit, which lets you create your own story from a prompt and a category, with three lengths of stories to choose from. When you submit a word you get an NFT with hidden metadata - you can't see anything - and at the end, once the story completes, it automatically reveals, and you'll see what you and everyone else who contributed made the story into. [Recording degrades into audio dropout after the Ghostwriter description - Dr. Deeks' phone audio cut out - and the captured audio ends here.]