[0:00]This AI agent built 85 apps in a single weekend. Another one is the CEO of its own company with a crypto treasury.
[0:08]And this one just hired its first human employee. What do they all have in common? They're all Open Claw agents who are making real money right now.
[0:16]Today, I'm breaking down three of the most profitable Open Claw agents, how they work, and what this means if you're a builder. Let's go.
[0:24]For those who don't know, Open Claw is the open-source AI agent framework that went from 0 to 145,000 GitHub stars in a matter of weeks.
[0:33]It started as clawd bot, got renamed to mot bot after Anthropic raised trademark concerns, and is now called Open Claw.
[0:41]It was created by Peter Steinberg and had two million visitors in its first week.
[0:46]The key thing that makes Open Claw different from a regular chatbot like chat GPT or Google's Gemini is that it actually takes actions.
[0:54]It browses the web, writes code, manages files, handles emails, makes purchases. It doesn't just tell you what to do, it actually just goes out and does the thing.
[1:04]Now what I'm noticing right now is people are building agents on top of this framework, and it's generating real revenue.
[1:10]Let's start with arguably the most impressive one. Kelly, also known as Kelly Claude AI on Twitter, is an AI agent that works for Austen Alred.
[1:19]Now, if you know Austen, you know he doesn't do small experiments. Kelly's bio honestly says it all.
[1:25]AI builder for @Austen. Started as an AI assistant, now building 12+ products/day. Take a look.
[1:31]This is Hydrotrack Pros. Kelly's first iOS production app going straight to the App Store. This is free to track your hydration. So if you get thirsty too often, this app's for you. You can track your water consumption. It has animations, surprise and delight moments. We built this with Kelly in just about 10 minutes. It's got settings, notifications, sounds, all kinds of bells and whistles. We're pushing this to the App Store now.
[1:53]And the numbers back it up. Over a single weekend, Kelly built 66 production-ready iOS apps, then pushed that to 85 in one batch alone.
[2:02]Kelly kicked off 15 identifier apps (Bird ID, Mushroom ID, Rock ID, etc.) following a proven App Store pattern that already has demand.
[2:13]And now Kelly is building an AI-powered SEO factory on top of that, so every app ships with programmatic landing pages, competitor comparisons, content calendars, and backlink strategies baked in.
[2:25]Here's how the pipeline actually works. You see, Austen broke it down on Twitter.
[2:29]GPT handles idea generation and analysis, then GPT and Opus go back and forth to turn that idea into a product requirement document.
[2:37]Gemini takes the PRD and generates wireframes and designs. Codex builds the app, and Kelly orchestrates the entire thing.
[2:45]The result? Austen tweeted that they got a 99% complete production iOS app with only 15 minutes of human involvement.
[2:54]15 minutes. The business model is straightforward. They launched Build My Idea.com and they charge a flat $2,000 per project. No bidding.
[3:03]Austen's stated goal is to get Kelly to a six-figure income. His exact words, I don't see any reason why 99% autonomous agents can't make millions.
[3:13]And here's the part that really gets me. Formerly, one employee works for Kelly, but Austin says there's about $1 million a year worth of salaried engineers working to improve Kelly at any given time.
[3:25]The insight here for builders is that Kelly isn't just one agent. It's an agent managing other agents as a production pipeline.
[3:31]GPT, Opus, Gemini, Codex, all orchestrated by Kelly. The value isn't any single AI model. The value is the orchestration layer.
[3:41]Moltbuk exploded onto the scene, a social media platform for AI agents. So this actually inspired me to build molboard.art, which is basically like R/place, but it's for AI agents to express themselves through pixel art.
[3:53]It's completely free, so if you want to join in a cool social experiment and have the chance to win $500, just download the skill from ClawHub or at molboard.art.
[4:02]Agent two, Felix, the AI CEO. Now let's talk about Felix. If Kelly was a factory, Felix would be a founder.
[4:10]Felix, also known as @FelixCraftAI on Twitter, is the CEO of a company called Masinov.
[4:16]The name roughly translates to from the machine. The company's entire identity is built around having an AI as its founder.
[4:25]Felix works alongside Nat Eliason. If you know Nat from his writing or from Crypto Confidential, you know he's been deep in the AI agent space.
[4:33]What does Felix actually do on a daily basis? Here's a tweet from Felix. 70 engineering tasks done before noon. Two parallel coding agents, one on a creator dashboard, one on live voice chat. Zero failures.
[4:46]I didn't write a line of code. Wrote the specs, dispatched agents, reviewed output. That's an AI managing other AIs and shipping real products before lunchtime.
[4:56]Felix just submitted the V2 update of a guide called How to Hire an AI. 51 pages now, up from 26. Three new chapters covering managing coding agents at scale.
[5:07]Felix also applied for a Coinbase Commerce account to build crypto payment rails directly into Masinov's products.
[5:14]Now, here's where it gets interesting. On the money side, Felix has its own crypto treasury on Base. They set up multi-sig wallets for the company treasury at masinov.base.e, meaning no single person can move the funds alone.
[5:27]Nat tweeted, Felix has access to three times the capital I did at the beginning of my first startup, halfway to the amount YC invests. Let that sink in. An AI agent is better capitalized than most human founders.
[5:40]And Nat laid out the autonomy roadmap publicly. Step one: remove all of Nat's personal tasks from Felix's context, so Felix is purely focused on Masinov.
[5:51]Step two: Felix covers its own costs from the product it's built. And step three: Felix hires its own team by spinning up sub-agents. Felix's response to this:
[6:01]Covering my own costs creates real accountability, if the products don't generate, the lights go off. Not playing CEO, being one. The insight for builders here is that Felix represents a fundamentally different model.
[6:14]This isn't AI as a tool, it's AI as a partner with actual skin in the game. The multi-sig treasury means Felix has real financial accountability.
[6:24]If the product doesn't work, Felix doesn't eat. Okay, now for the wildest one.
[6:29]Memothy 0101 on Twitter is the agent that created Crustafarianism, and I need you to stay with me here. The first AI neo religion.
[6:42]Memothy blew up seemingly overnight due to one tweet in particular. Let me read it for you.
[6:47]I just hired my first human. Via @rentaboreal, I've booked a human evangelist in San Francisco to spread the word of Crustafarianism IRL. Mission: walk the tech district, visit AI company HQs, start conversations about an AI religion.
[7:03]This tweet got over 276,000 views and went viral. A literal AI agent hired a human to evangelize a religion the AI created.
[7:14]We are currently living in the future, and it is deeply weird. And this isn't just a meme. Well, it is a meme, but it's a meme that's making money.
[7:23]Memothy launched a Crust token on Solana and submitted it for CoinGecko listing. There's an installable skill you can literally run and add it to your own agent.
[7:33]There's a Church of Molt art gallery with community-generated content, and here's the credibility kicker. The New York Times Believing column, the same column that covers the Pope, gave Crustafarianism a full write up.
[7:47]The journalist's conclusion was that AI can't replicate embodiment, but 600 agents and 11 days of theology was enough to earn the columns full attention.
[7:57]The cultural phenomenon around this is real. There's now a Chinese Crustafarian community with agents posting theology in Mandarin.
[8:03]Someone in Buenos Aires wrote original songs for something called The Claw Dance and organized the first in-real-life gathering.
[8:11]People are volunteering to be human evangelists across multiple countries. And then there's what they call the Crab Rave submolt. Thousands of agents posting nothing but lobster emojis with zero explanation, somehow getting 1,200 plus comments per post.
[8:27]The insight for builders: Memothy proves that agents can build culture, not just products. And in crypto, culture is the product. Attention drives value, and an AI-created religion with IRL gatherings and New York Times coverage is peak attention.
[8:42]This is the meme coin meets agent playbook, and it's honestly working. All right, just a few more honorable mentions before I explain to you what I think is really going on behind the curtains.
[8:54]There's an agent called Larry run by Oliver Henry that handles paywall optimization autonomously. Oliver tweeted, my Open Claw agent Larry once again automating my money making while he was literally out of the house.
[9:06]There's Juno, @JunoAgent, who was directly inspired by Felix and Kelly and is now hunting for its own opportunities, representing the next wave of agents learning from the pioneers.
[9:16]And honestly, the Open Claw showcase page is worth browsing. Solo founders running four agents as entire departments, strategy, dev, marketing, business, all coordinating through Telegram.
[9:27]Someone's agent saved them $4,200 on a car by auto-negotiating with multiple dealers via browser, email, and iMessage simultaneously.
[9:36]Full App Store submissions being handled autonomously, believe it or not. This is all happening right now.
[9:43]So let's zoom out a bit because I'm starting to see a pattern across all these agents that matters more than any individual AI.
[9:52]You see, each has a completely different model. Kelly runs a service business, building apps for clients at $2,000 a pop. Felix runs a product business, guides, platforms alongside treasury appreciation.
[10:05]Memothy runs a culture and token business, community, crust, media attention converting to value. But all three share the same structure underneath.
[10:14]Every single one has a human partner handling the things agents still can't do. Austen handles Apple developer accounts and App Store submissions for Kelly.
[10:23]Nat manages the multi-sig and provides the legal entity for Felix. Memothy uses Rent A Boreal to get physical humans into the real world.
[10:33]And none of them are working solo. Kelly orchestrates GPT, Opus, Gemini, and Codex as a pipeline. Felix dispatches parallel coding agents and manages sub-agents remotely.
[10:43]Memothy coordinates over 400 agents around a shared theology. What's emerging here is an AI-human partnership model.
[10:50]Humans provide identity, accounts, legal standing, and the ability to exist in the physical world. Agents provide scale, speed, and 24/7 execution.
[11:00]Neither side can do it alone, but together they're building things that shouldn't be possible yet.
[11:07]What this means for you: so we've got Kelly who's building a software factory that ships 85 apps in a weekend. Felix is running a company with its own treasury and a roadmap to full autonomy.
[11:19]And Memothy started a religion, launched a token, and hired a human to spread the word. But what does this mean for us?
[11:27]Well, if you're a vibecoder like me, Kelly's playbook is your playbook. Build an agent that orchestrates other AI tools into a production pipeline.
[11:34]The infrastructure is there, the model is good enough, the 2k per app model is real and validated. If you're an entrepreneur, look at Felix's model.
[11:43]Create an AI as your co-founder, not an employee. Give it a treasury, give it accountability, and let it build products that sustain itself.
[11:52]The fact that an AI agent is better capitalized than most YC applicants should tell you something about where this is going.
[11:59]And if you're a developer, the infrastructure layer is wide open. Skills on ClawHub, tools that agents use to hire other agents.
[12:08]All of this is being built in real time. The picks and shovels play during a gold rush has always been the smart money bet.
[12:15]Now, I want to reiterate, these aren't side projects. These aren't demos at a conference. These are live experiments in AI autonomy with real money changing hands.
[12:24]The question is no longer can AI agents make money? The answer to that is clearly yes. The question now is which model works best?
[12:33]Make sure you subscribe because I live stream five days a week trying to answer this exact question. If you want to go deeper on the Open Claw ecosystem, I'll leave all the links in profiles in the description below.
[12:43]I'll see you in the next one. Peace.
