I Built an Entire App in Only 52 Min with Codex & Idea Browser

Go to https://startup-ideas-pod.link/idea-browser to get validated startup ideas every single day and prompts to build your big idea Join me as I chat with Alex Finn about going from no idea to building a functional app using AI tools. We use Idea Browser to find trending business opportunities, develop a wellness habit tracking app concept with ChatGPT, and build it using Codex's AI agents. The episode showcases how one person can leverage AI to find and build an idea into a product. Get the blueprint to go from idea to an app in one afternoon: https://startup-ideas-pod.link/ai-builder Timestamps: 00:00 - Introduction and overview 00:39 - Finding app ideas with Idea Browser 05:19 - Using ChatGPT to develop the app concept 17:36 - Overview of Codex and its capabilities 22:25 - Building the MVP while spinning up multiple AI agents 46:04 - Marketing strategy for launching the app Key Points: • Using Idea Browser to find validated business ideas by analyzing trends and market insights from social platforms • Leveraging ChatGPT to develop a business concept into a detailed Product Requirements Document (PRD) • Building an app with OpenAI's Codex using both local and cloud agents simultaneously • Implementing a marketing strategy through organic social media outreach The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ Boringmarketing - Vibe Marketing for Companies: boringmarketing.com The Vibe Marketer - Join the Community and Learn: thevibemarketer.com Startup Empire - get your free builders toolkit to build cashflowing business - https://startup-ideas-pod.link/startup-empire-toolkit Become a member - https://startup-ideas-pod.link/startup-empire FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/ FIND ALEX ON SOCIAL Youtube: https://www.youtube.com/@AlexFinnOfficial/videos X/Twitter: https://x.com/AlexFinnX Creator Buddy: https://www.creatorbuddy.io/

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Greg Isenberg?

  Alex, by the end of this episode, what are we gonna learn?

• 0:03Speaker 01

  You are gonna learn how you can go from no idea whatsoever to brilliant idea to having the idea fully built out with AI agents.

• 0:12Greg Isenberg?

  Okay, and could you give us a sneak peek of some of the tools that we're gonna be using?

• 0:18Speaker 01

  We're gonna be using two really awesome tools. We're gonna be using Idea Browser to come up with a really amazing business idea, and then we're gonna be using OpenAI's Codex to build out that idea.

• 0:30Greg Isenberg?

  Okay, let's get started.

• 0:39Greg Isenberg?

  I must confess that I built Idea Browser for myself. So I am a co-founder of Idea Browser, but I think that if it's valuable for me, it's gonna be valuable for other people. So let's get to it.

• 0:52Speaker 01

  And I must confess, I have absolutely no affiliation with Idea Browser whatsoever. I just find it awesome to find ideas. So let's do it.

• 0:59Greg Isenberg?

  All right.

• 1:00Speaker 01

  I am going to share my screen. First of all, props Greg on Idea Browser. Awesome, awesome platform for me.

• 1:09Greg Isenberg?

  Well, the context there, by the way, is that like I run a holding company. So we're constantly incubating, investing and acquiring businesses. So we need ideas. Ideas are the lifeblood of our business. So the whole idea is like, if you can scrape Facebook groups, subreddits, get data and create validated ideas, then you have an unfair advantage.

• 1:31Speaker 01

  Yeah, and I, you know, based, I talk to a lot of vibe coders and new people to AI every single day. And, you know, building has never been easier. The challenge is the ideas. People can't come up with good ideas. And so this is a fantastic way to find them. So you can find things to build out. I personally use this kind of as like a boxing gym in a way, Idea Browser, where it's like, okay, I'm gonna go into this. I'm gonna find ideas. I'm gonna build them out. Not always necessarily will they be million dollar businesses, but at least I'll get that practice of going from my idea to building. And so what I'll do is kind of walk you through how I use it and we'll choose an idea and we'll start building it out. But where I like to start off Is market insights. What's cool about market insights is it pulls from a lot of the big

• 2:14Speaker 01

  social media networks, right? Reddit, Facebook, and you can see what people are talking about, see their pain points, see where there might be solutions and maybe come up with ideas from these. So I don't know too much about psychedelic assisted therapy, but I know a little bit about email deliverability. And so I'll come in here, I'll see what pain points people are having with email deliverability, sudden drops in deliverability. If you have a newsletter like I do, you know, this is something you run into all the time. Fear of Google Outlook policy changes. You can really find some cool, interesting pain points to solve from these marketing insights. They give you some solution gaps, places where there are no solutions at the moment, segments that are

• 2:59Speaker 01

  underserved, a lot of really cool ways you can pull ideas from here. But where I seem to personally find the most ideas to build out is I like to go to trends. This is just like purpose built, ready, served on a platter, things you can build out. Like, I don't even have to think too hard, right? Like, website designed for small businesses, wellness app, this is a cool one. So this is, this is one I've been thinking about, and it looks like it is having a lot of volume and growth at the moment. And so what I'll do is I'll look for a trend that I myself have interest in, because the best products to build out are things you have interest in. I'll click in,

• 3:37Speaker 01

  and what this is going to do is give me more information on this trend. So I can actually pull up the wellness app here. It'll tell me a little bit about how many monthly searches this app or this trend is getting, tell me about what people are searching for with this trend. A wellness app refers to a digital platform designed to support individuals in achieving and maintaining their health and wellbeing goals, and gives me a little bit info on how many people are searching for this. And as you can see, a lot of people searching, not a ton of competition. So this, for me, when I'm coming into Idea Browser, and I'm looking for ideas to build out and practice my building skills, this is kind of exactly what I'm looking for.

• 4:17Speaker 01

  Things that are having good growth, things with not a lot of competition, and things where I myself am interested. I'm interested wellness, I work out, I try to eat relatively healthy, not including chocolate. And so this is something I, for me, is a really good candidate for to build out.

• 4:38Greg Isenberg?

  It's funny because like when you put it like that, find something that is in a growing industry with low competition that you personally understand and love, like obviously that you know, your chances of success is going to be highest there.

• 4:52Speaker 01

  Well, the chances of success honestly are 100%. The reason being is like, at the very least, even if no one buys the product at all, you've built a product for yourself, right? That'll improve your own life. And like for me, my business creator buddy, I built it because I wanted to build an AI that analyzes my posts. At the very least, if no one bought it, I would have a product for myself to improve my own content. And so that's why it's always a win if you build things you're interested in. So here's my workflow. This is what I do. You know, I got into trends. I found a trend I really like. I have a prompt I built called the idea browser prompt, purpose built for idea browser.

• 5:32Speaker 01

  Where I'll go, I'll take this trend, I'll copy it, I'll go into chat GPT, which for me at the moment, GPT five is the most powerful model there is. Sonnet four or five incredible for coding, the best for coding, but for straight up for chatting, ideating, I'm still a GPT five guy. And so what I'll do is I will take my prompt, and we can put this in the description down below. But basically, the prompt is this, I'm giving you a link to a business idea from idea browser. Your job is to act as a professional app builder and creative strategist. Read the page and then generate three high quality, unique app ideas inspired by that idea. For each app idea include the name, a concept summary,

• 6:17Speaker 01

  core features, why it works, and why it would succeed in today's market. Keep it practical and make it easy for a solo founder to build out. So I have that prompt. I'll take our trend we found. I'll put it in here. So here's the idea, paste it in. And what this is going to do is have chat GPT, go into that website, see what that trend is all about, and then give us a few ideas for what we can build out in our next stage, which will be going into codex and actually building out that idea.

• 6:55Speaker 01

  So we have a few different ideas here. We have Mood Mosaic, which is an expressive mood journaling and creative wellness app for people who want a more visual playful way to track emotional health. Rather than writing pages of text, users paint their mood via simple sketches, colors, stickers and stamps. Maybe, a little out there. WellSync Buddy. WellSync Buddy is a wellness accountability and habit pairing app that helps pairs, small groups, synchronize and support each other's wellness routines.

• 7:27Speaker 01

  And it's for people who thrive on social accountability. That's actually a multiplayer social network wellness app. It's not a bad idea. And then we have BioRhythm Coach. BioRhythm Coach is a personalized micro wellness app built around users' individual daily bio rhythms. It nudges users to schedule task breaks and wellness micro practices, like stretching, breathing and light exposure, in alignment with their energy peaks and troughs. That one's pretty cool. I think, I don't think any of these are bullseye for what I want to build, but what we can do from here is we can take certain ideas and combine them into one. Right? Like, some of these are a little abstract, like Mood Mosaic. I don't

• 8:11Speaker 01

  really want to build like a drawing app. WellSync, building a multiplayer social network's a little tough for one podcast. BioRhythm, I kind of like getting daily habits based on your goals. So let's do this. I think what we can do is we can say, "I like idea number three, but it's a little abstract. Build me a plan for a habit tracker that also builds us wellness routines."

• 8:50Speaker 01

  So what I really liked in this one was the wellness, micro wellness app. I like that idea, being able to track every little wellness thing you do and getting recommendations around how to be more well. So let's, we can go back and forth with chat GPT here and get more ideas around that.

• 9:06Greg Isenberg?

  Yeah, I like that idea. Like we have to do lists for our professional lives, but we don't really have to do lists for our longevity lives.

• 9:15Speaker 01

  We don't, we don't. And I think a lot of reason why is people just don't think of ways to be well. So putting a little bit of AI spin there where it recommends those things and you make a to do list out of that. I think that could be cool.

• 9:29Speaker 01

  So let's see what we got here. Concept overview habit flow is a dynamic wellness habit tracker that doesn't just track what you do. It designs and adapts your daily wellness routines automatically based on your energy levels, goals and feedback. The app blends traditional habit tracking with adaptive scheduling and AI coaching help users build sustainable personalized wellness systems. This is something I'd use. Greg, would you use this?

• 9:53Greg Isenberg?

  I totally would. I'm actually, and I'm like in a health kick right now. I'm like actively trying to, to be more healthy and like I need it in my face a little bit more. So yes, I would.

• 10:05Speaker 01

  Even gives you the monetization. So free tier, how you can make money off of it, upsell. We really, at this point, have not done any sort of creative thinking. We've leaned on idea browser and it's just kind of fed us ideas and the AI helped us come up with the winning idea.

• 10:22Greg Isenberg?

  I mean, there's even a lazier way to do this. We'll do, you know, for another video, but like the lazier way to use idea browser, if you go back to it for a second, like you can, if you see the related idea next to the trend on the right hand side on the right bar, there are, and click, click that idea.

• 10:43Greg Isenberg?

  So in Idea Browser, there's almost a thousand, I think, ideas and based on all these trends. So there are pre-cooked ideas based on, you know, and these are curated by human beings. So it's like, we go through them and we're like, hey, this is good or bad. So you're right, you're right, you can use ChatGPT, you can use Claude, you can use some of these tools to go and ideate, or you can just use the pre, the database. And there's one more tool that you can also use that I should just make mention is if you go to, yeah, Tools,

• 11:20Greg Isenberg?

  Research Your Ideas. So we won't have time to do this now, but let's say you picked an idea you really liked. Like, we could have picked, like, that one, it sounds like we have that idea that we liked from ChatGPT. You can actually put it into Idea Browser. It's gonna, it takes time, it takes up to 24 hours, but it actually generates a report on exactly, like, pretty much step by step, how you should be building this business.

• 11:46Speaker 01

  So it's like a, like a deep research tool.

• 11:48Greg Isenberg?

  Deep research, deep research for ideas, yeah.

• 11:51Speaker 01

  That's sick.

• 11:52Greg Isenberg?

  And it's cool seeing how other people use it, by the way, like, I, we see people use, some people want, it's just like more of a preference thing, you know, some people like pasta, some people like pizza, some people like relying on the trends and just like, kind of going with it, and some people just say, I like human beings giving my ideas and that's why I'm just gonna go to the database.

• 12:16Speaker 01

  Yeah, I mean, I think the key is this, is not to overthink, and I think the main problem a lot of people have is they kind of wait for the perfect idea, and then the perfect idea never comes and they end up accomplishing zilch, right? What, the key and how I kind of got to my winning business idea was I just kept building things, I just kept choosing ideas, and you learn from each idea you build out, oh, this is actually interesting, I like building this app, and you eventually get to where you want to go. And that was what was great about Idea Browser, you have a trillion ideas in front of you, just choose one, start building it out, and If it's a losing idea, you still win because you figure out what the next idea you want to build out is and you get closer to where you want to be. So just build something out.

• 13:01Greg Isenberg?

  Totally. I mean, even the idea for Twitter, for example, came from Odeo, like the original concept. I don't know if you know this story.

• 13:09Speaker 01

  I don't.

• 13:10Greg Isenberg?

  Yeah. So Jack Dorsey raised a few million dollars to build out Odeo, which is a micro podcasting app idea. And he built it and started to just, you know, I guess it was too early to do that. And he was kind of like, let's just make a text and came out, you know, basically came out with the idea from for Twitter through that process. Same is true with Instagram. You know the story of Instagram, how it started off as Bourbon?

• 13:35Speaker 01

  No.

• 13:37Greg Isenberg?

  Bourbon was a location. It was kind of like Foursquare, actually. Remember check-ins?

• 13:44Speaker 01

  Yes.

• 13:44Greg Isenberg?

  Like checking in. Yes. You would check into places and you would post photos of the place you checked in on. What they learned was people just cared about the photos. And so they're like, let's just build a photo sharing app. So your best idea usually isn't your first idea, but it kind of gets you there to your point.

• 14:07Speaker 01

  And I think it's the same thing with OpenAI. Isn't it the same exact story where it wasn't a chatbot at all?

• 14:13Greg Isenberg?

  Right.

• 14:13Speaker 01

  They just kind of demoed it as kind of like a proof of concept. Hey, what would happen if this was a chatbot? And it turned out everyone using OpenAI was just doing the chatbot like, okay, we'll turn this into a chatbot product. And they released GPT 3.5. And like, that's all they've been doing since is building on top of that chatbot. So that's how you come up with your trillion dollar idea. You just build something and you'll along the way, find what the right thing to do is.

• 14:39Greg Isenberg?

  All right, let's go back.

• 14:42Speaker 01

  So let's do this. We've kind of ironed out our idea. The last thing we need to do before we start building, which this will be the fun part, is we want to turn this into a PRD. So a product requirements document. AI agents are really, really good when they have a detailed plan, when they have a product requirements doc to work off of. So we want to build this and then hand it off to our Codex agent who will start building the app for us. And so what we're going to do here is we are going to say, okay, I really like this idea.

• 15:17Speaker 01

  Let's turn the MVP very simple version

• 15:27Speaker 01

  of this app. And for those who don't know at home, minimally viable product. So like kind of like the first version, the app that is usable, that gives value to the user, very simple version of the app into a PRD, we can hand off to an AI coding agent.

• 15:46Speaker 01

  And I'm going to hit enter on that. And now ChatGPT 5 will build us the PRD for Habit Flow. Actually, pretty good name, if you ask me. And we were going to take this PRD. And in a second, I'll show you how we hand it to Codex to start building some cool stuff out.

• 16:04Greg Isenberg?

  I gotta be honest, I haven't like played around with Codex all that much. Am I making a mistake?

• 16:11Speaker 01

  You are making a mistake. Now, do I think it's the end all be all? You'll never have to use another AI coding tool again. No, I think it works really, really well in complement to like a Claude code, for instance.

• 16:23Greg Isenberg?

  Yeah.

• 16:24Speaker 01

  But there are some features and functionality, which I'll show you when we hop in, that you can't do anywhere else. Primarily their concept of handoff and cloud agents, which basically gives you the ability to spin up tasks wherever you are, whenever you are, cross-platform, and constantly be productive, no matter what you're doing.

• 16:47Speaker 01

  So let's do this. Let's take this PRD. Actually, let's do one more thing here. Can you make this a markdown file, please? This is the way AI agents read, is in markdown files. So boom, perfect. It's going to spin this up into a markdown file. And in the meantime, what I'm going to do is I'm going to switch over to Codex. So for those who want to work with us at home alongside us, which I think is a very smart idea, what I have here is Visual Studio Code, completely free. You can download it, no monthly fees or anything. And I installed the Codex plugin for Visual Studio Code, which is going to give you this

• 17:30Speaker 01

  nice Codex UI here over on the left-hand side that we can work in.

• 17:36Speaker 01

  But the first thing we're going to do is we're going to create a new file. I created a new project in Visual Studio Code. I'm going to create a project dot M D, a Markdown file. And I copy over our new P R D here. It's going to have everything about the project. I'm going to hit save. And now what we can do is switch to Codex. For those new to Codex, this is a plugin for Visual Studio Code. So you can just install it. It connects to your Chat G P T account. So as long as you're paying the $20 a month or whatever for the kind of the base tier, you'll have access to Codex. Now, there's a couple of things different here if you're more used to Claude Code or Cursor.

• 18:20Speaker 01

  With Codex, you have two options.

• 18:24Speaker 01

  You can work locally or you can work in the cloud. So working locally is kind of probably what you're used to with Claude Code, probably what you're used to with Cursor where it just sees the code in your in your I D and edits it, writes code, changes code, whatever. The difference here is you can run this in the cloud. And basically what that means is you can spin up A I agents that work in the cloud, on the Internet and does things for you asynchronously, writes code for you asynchronously. And the advantage to this is you can now spin up these agents from your I D plugin here, like you see with Codex, from your Chat G P T app on your iPhone.

• 19:08Speaker 01

  You can spin up agents from Codex on the web. So if you go to chat G P D dot com slash Codex, you can spin up agents from your web browser. You can spin up agents to do tasks for you wherever you are, mobile browser, I D, and they all work together, hand off tasks to each other, which allows you to basically unlock productivity while you sleep, while you're on the go, while you're at the gym, wherever you are. You can have A I agents work for you, building out code and doing whatever you want.

• 19:41Greg Isenberg?

  That's a big deal.

• 19:43Speaker 01

  It's a really big deal. It's a, it's a, it's a whole new paradigm of building that we really haven't had before where doesn't matter where you are, doesn't matter what you're doing, you can have code being built. My personal favorite way of using it is before I go to bed every night, I challenge myself, I need to spin up three tasks for AI agents. So before I go to bed, before I turn off my computer, I have to give Codex three tasks. So whether it's implementing features, whether it's doing other tasks, which I'll show you in a second, whether it's marketing, product, whatever, just having three tasks spun up before I go to bed so that when I wake up, my app has advanced. It's a really, really cool new way to be able to use AI agents.

• 20:27Speaker 01

  With that being said, I'll show you all the other cool stuff you can do with cloud agents in a second, but let's get the kind of base app set up here. So I'm going to say work locally. And I'm going to say, I created a PRD in project dot M D. Please build out the MVP using NextJS

• 20:56Speaker 01

  and Superbase. So that's going to be our tech stack. NextJS for the front end, Superbase for the database. You can choose your model. You have many different options here, high, medium or low. I've been using Codex high for everything, but I just watched the Codex Product Manager build out an app live at the OpenAI Dev Day and he used low for everything. I was kind of shocked. So I'm going to use low for this. And I'm going to hit send and it's going to start building out our MVP.

• 21:28Greg Isenberg?

  You're like, if it's good enough for him, it's good enough for me.

• 21:31Speaker 01

  Yeah, I was just like, by default, I want the smartest model. And so I've been using high and which takes forever, takes a long time. But the head guy who did like a live demo, building out a complex app for the entire convention used just low for everything. I'm like, okay, I guess low is good enough.

• 21:47Greg Isenberg?

  It takes, it takes longer and doesn't it take more tokens?

• 21:52Speaker 01

  Takes more tokens, takes longer, but I don't, my mindset's always been like, I just want the smartest employee possible doing everything for me.

• 21:58Greg Isenberg?

  Right. You're putting in like the, the most premium gasoline.

• 22:03Speaker 01

  Exactly. I want the most premium everything. If I'm drinking water, I want to drink mountain valley water. I can't drink out of a tap. Are you kidding me? Attaboy.

• 22:11Greg Isenberg?

  There we go.

• 22:13Speaker 01

  You know, this is, this is the GPT high of water, but apparently you can do the GPT low tap water and it still works well.

• 22:23Speaker 01

  So that's building that out. Now, while that's building that out, I'm going to show you the power of Codex. And for the record, OpenAI is not paying me, but also for the record, I will happily take an OpenAI check. I'm going to switch as this is working to the cloud. So run in the cloud. Actually, let's do a new chat here. I'm going to say run in the cloud. And I'm going to say, take a look at project.md. Make sure we're on the right. Yep, we're on the right repo here. So you're going to connect this to your GitHub as well. We're on the wellness repo. And build us an entire marketing plan for this app.

• 23:10Speaker 01

  And so now what we're doing with Codex is we have the local agent building us out the app, the MVP. And now I'm going to spin up a cloud agent to build us a marketing plan on the cloud for the app. And I hit send on that.

• 23:26Speaker 01

  And that is the power of Codex and like these cloud agents that you really didn't have before in cursor or cloud code that you can now take advantage of where I'm now spinning up AI employees in the cloud to do other tasks while the local agent is building out our app for us.

• 23:44Greg Isenberg?

  It's kind of crazy because if I had downloaded VS Code, downloaded Codex, I wouldn't have known that, like just by looking at local versus cloud.

• 23:56Speaker 01

  Yeah, I mean, it's kind of a new paradigm of building. Also too, I think these companies can do a little bit better of a job of explaining to people how everything works. But yeah, it's a whole new way of thinking about AI coding. It's instead of thinking about it like I'm like pair programming with an intern like you would with Claude Code or Cursor. It's now I'm kind of the CEO of a company and I'm hiring employees to do things for me as I work, right? I have my software engineering intern building now. I in my chat have in the cloud, my marketing head of marketing going and building this out and building out a marketing plan. And you can just spin up agents as you go in the cloud to do a lot of things at once.

• 24:44Speaker 01

  And I can actually show you here as this is going. Oh, you can actually see it right here. So you can see the work it's doing in the cloud. Here's an entire marketing plan for habit flow. Objectives and KPIs we can track, target audience and who we're going to be targeting, resilient achievers, 25 to 40 year old professionals like us. It's exactly, exactly what we were thinking with this. Positioning messaging pillars, competitive landscape. It did, look, it did a research on competitors for us, did everything we wanted as our other agent was working on us and actually building out the MVP. And it looks like that is still working here as well as 1,200 lines. Because we had 1,200 lines of code built out for us while our marketing agent was working the cloud to build out that plan.

• 25:30Greg Isenberg?

  While we were talking smack about OpenAI, 1,200 lines that they're just beautiful lines that they're creating for us.

• 25:38Speaker 01

  And what's really cool is if I actually, it's probably not going to be too visible on the camera. But if I go into Codex on my phone and I refresh here on my chat GPT app, I have that entire marketing plan on my phone as well.

• 25:56Greg Isenberg?

  That's cool.

• 25:57Speaker 01

  Right? And so it's completely cross-platform. Everything I do in the ID is in my mobile app. Everything I do in my mobile app is on the browser and vice versa across the board.

• 26:09Speaker 01

  Which is, you know, an ecosystem I don't think any other AI company can really pull off right now at the moment, which is really awesome.

• 26:20Speaker 01

  So we have it still building out our app. We're gonna have the MVP done when it does that. But like, I think typically when people are building out apps like this, during this time when the agent's building, they'll go on X, they'll doom scroll, they'll do whatever. All of a sudden, two hours gone, they got nothing done. If you kind of shift your mindset now to having cloud employees, AI employees, which is unlocked with Codex, you can get a lot more done in these kind of in-between moments, right? Like, while this is working, I can maybe spin up a new chat, make sure we're working the cloud, and we can say,

• 26:58Speaker 01

  "You're a product manager. Come up with a roadmap for the app based on project.md." And I hit send. Now I have my product manager employee working for me, all while I wait for the MVP to be built out, which looks like that is actually good to go and set here. Boom. Okay, perfect. Yep. So now we can run this. And we can test it out. So let's see, we didn't set up the database just yet. So this might not work right away, but if this doesn't work right away, we can run through setting up the database as well. But I'm going to, let's see here.

• 27:42Speaker 01

  I'm going to run this. npm run dev. Let's see. Run. Oh, we gotta run npm install. Okay. Install. So this is going to install all the dependencies, all the tech, all that, so it's good to go. And then we'll be able to run the app after this. All right, so we'll do a little bit of debugging here. We're getting errors trying to install the dependencies. I'm gonna go in, I'm gonna go to Codex. I'm getting errors when running npm install. Here they are.

• 28:20Speaker 01

  Let's make sure we run this locally. Yep, we're running it locally. Hit send. And it'll go through and it'll figure out the errors and fix whatever needs to be fixed here.

• 28:30Greg Isenberg?

  I'm happy that there's errors, to be honest, because, you know, it happens, right? It's a part of the process.

• 28:37Speaker 01

  It looks like, yeah, it looks like it had the wrong version of Superbase in the, uh, in the dependency. So it was trying to install a version of Superbase that doesn't exist, which is an issue. So let's try that again.

• 28:53Speaker 01

  Still having errors with Superbase. Let's do this. Let's just say, for the sake of this, seems to still have issues with Superbase for this MVP. Let's just save data locally.

• 29:12Speaker 01

  I'm going to hit send so we can get a working version first and then we'll implement Superbase right after so we can get that working off the rip.

• 29:19Speaker 01

  So that's going to kind of remove that dependency that's giving us fits here so we can see, uh, kind of the MVP version here that we can test.

• 29:29Speaker 01

  But again, unlike other apps, what we can do with Codex is we can have many different strands of work going on at the same time. So we still have our roadmap going here. So we have the roadmap. Let's check in on that.

• 29:45Speaker 01

  So here's our roadmap. Vision: Deliver a delightful adaptive habit tracker that keeps people consistent in wellness routines. So guiding initiatives, product alignment, phase one, the core MVP experience, which we're working on.

• 29:59Speaker 01

  Improved habit here. It's via smart nudges. Let's see, phase three, expansion and monetization experiments. So it really builds us out step by step. What we're going to be building out with our product manager, uh, agent here, which really cool. You can always make sure you're being productive.

• 30:19Speaker 01

  We can go back to our local agent who is fixing our dependencies here.

• 30:25Speaker 01

  What are you using at the moment to kind of vibe code? Are you using Claude or what have you been playing with?

• 30:31Greg Isenberg?

  Yeah, I mean, I've been in a Claude code phase for maybe, what has it been, three months?

• 30:39Greg Isenberg?

  But have been seeing, it's kind of one of the reasons why I wanted to have you on, like have been seeing some of the Codex stuff. And I'm starting to get tempted to, you know, to go deeper into it.

• 30:55Speaker 01

  It's, it's worth playing around with to see how it fits into your workflow, right? Because it works a little different. I find in some ways Codex isn't as proactive. It seems like Claude always does a little bit extra of what you ask, which some people annoys them. Some people they love. Codex is almost the opposite where you have to nudge it along sometimes to do things. So it's worth playing around with to see if it's compatible, right? Like the same way you'd interview employees to see if they work well in your company. It's like interviewing the AI agent.

• 31:28Greg Isenberg?

  Yeah, you also get to know, you know, who they are and how they work. And you can kind of, what might be good for you might not be good for someone else.

• 31:40Speaker 01

  Exactly, exactly. You know, there's people who live and die by cursor, right? And they all work in different ways and have different features and functionality. For me, I'm having, I'm, I'm having a lot of productivity, almost like 60/40 Codex to Claude code. Where, you know, I have the workflow I'm showing you here. And then if it gets stuck, I'll have Claude code kind of be the surgical go in and fix that like micro issue. That's, it's been working well for me.

• 32:11Speaker 01

  So I'm going to try installing this again. We change this around. Boom. Okay. So we added the packages. Looks like it installed. Let me see if I can do npm run dev and get this working. Looks like it's working. Okay, let's test this out. All right, I am going to switch over. Habit Flow. Build sustainable routines with energy aware planning. Adaptive wellness routines without the rigidity. Habit Flow calibrates your day every morning. Set up your focus, pick your starter habits, and let the app shape a plan that responds to how you feel. Track completions, watch streaks grow, and help the momentum with gentle, data-driven nudges. Okay, let's see what, see if we got, oh, wow. So we have an entire onboarding flow.

• 32:54Speaker 01

  And again, like, first of all, this is awesome, right? Like, we have, like, obviously, this isn't the full app, but this is a pretty cool MVP, where we have, like, the intro, the onboarding experience, whole lot of other cool things in here. In this, really, you get these types of awesome results when you do things like build out a PRD ahead of time, right? When you develop your idea. And it takes a little bit of extra time to do, right, to go into Idea Browser, find a really good idea, develop it with AI. But you're going to save time in the long run, because you're going to be a lot more focused, and get better results from your agents, because you have that in-depth PRD built out. So let's play around with this

• 33:37Speaker 01

  a little bit, and see what we want to add, what we want to tinker with. And I can kind of show you how I use Codex to tinker with different things, and add on new features, and things like that. So it's personalized habit flow. Alex, check-in preference, morning check-in. For me, focus is the issue. You get so many different AI tools out there, I can never focus on one thing at a time. I'm sure you got that same issue. Starter habits, 10-minute mobility flow, 2-minute breath reset, I could use that. Hydrate, I'm pretty good at that. Let's get rid of that. Let's do, I could, I could do more deep work, sprints.

• 34:12Speaker 01

  Habits, I want to meditate more. Everyone talks about meditation. I want to meditate more. Let's do that. Let's add that to the list. Oh, okay. So it saves the habits we want to do and then we can save and continue. Please add your name. I gotta add my name in here. Alex. Save. You can hop into today to generate your first adaptive plan. Okay. So let's go to today. Let's see how that works. Today's pacing, tune your energy, then work through the tailored plan. How's your energy? I had a lot of coffee today, so we're gonna do five. Alright, we got an error. Let's, we'll do some debugging in a second.

• 34:48Speaker 01

  Plan mode, so it gives us, okay, it gives us some, some uh to-dos for wellness techniques we can do. And then routine logic primary rank habits. Okay, so this is pretty cool. We give it our goals and it gives us a bunch of to-do items to stay well. I like that a lot. Let's do a little debugging. Let's see what this error is. Set log is not defined. So a lot of people, they have really complex

• 35:15Speaker 01

  workflows when it comes to debugging and errors. They have MCPs that has Codex or Claude Code, watch your log and like screenshot your browser and this and that. They have all these. At the end of the day, they're probably usually trying to sell you something to get you to do these really complex workflows for no reason. Here's my really complex workflow for debugging. I highlight the error. I do command C

• 35:42Speaker 01

  and then I go into my IDE and I hit command V. And then I hit enter. I don't find any need to set up MCPs for this. I don't need to do any sort of advanced workflows. Please solve this error.

• 36:02Speaker 01

  And then I hit enter. This seems to work best for me. I don't need to have Codex connected to my browser or anything like that. I just copy and paste errors and it seems to fix it pretty well. And it looks like it's fixing the error. And at the same time, while it fixes the error, again, you probably want to get into kind of that Codex mindset of, okay, how can I be productive while my agents work? I'll spin up a new chat. I'll make sure I'm in the cloud and I can say, set up an AI chat interface where I can chat with a

• 36:40Speaker 01

  Wellness expert AI and get advice. So I thought of another feature to build out while this works. I'm going to give it to our cloud agent. And what I'll show you towards the end of this is how we can start pulling down these changes and implementing it into our app as we go. And so the key here, especially when you're using a really awesome tool like Codex is productivity, focus, and because we have these cloud agents, we can always make sure we have things cooking, things baking in the backend and have new features pumping out as we go. So let's see here, let's go into, uh, let's see, did we have this here? Let's go.

• 37:22Speaker 01

  The only downside is sometimes you get tripped up and you forget which agents are working on what. Here we go. Looks like it fixed the error. Let's go back in, make sure we're not getting an error on the screen anymore. Is this like kind of percolating ideas for you, Greg, around Codex and things you can do with it?

• 37:42Greg Isenberg?

  Yeah, totally. I also think, first of all, I agree with the whole MCP complicated workflow to the errors. So I've tried using some of those tools and it's, oh man, it's like, it just feels like I'm over engineering things. So I'm happy you said that. So that's, that was actually what I was thinking about. The second thing is, I was literally, I was thinking about this app and I, you know, one of the beauties about building an app, instead of just like sketching it, uh, you know, on a notepad, is it got me thinking that this in a mobile setting would actually probably rip. You know, not that we're going to build an iOS app today, but you know, maybe if people want in the future, we can.

• 38:24Speaker 01

  Follow up, follow up iOS focus on where we go. We show you how to build an iOS app. I like that.

• 38:29Greg Isenberg?

  Yeah, no, but I think that's, that's how I would think about it. And then what I do is I would, I would probably write out that idea and be like, hey, here's what, here's what I've built out. You know, this is, here's, you know, here's my vision for an iOS app. And I would use Idea Browser's AI agent to basically scan, you know, scan social sentiment, give data around, you know, is this a good idea and how, how would it implement it? So that's something that always exists. That is a pro feature, but. You know, that is, that's something I, it's like, you have to build it to get those ideas. You know what I mean? So.

• 39:13Speaker 01

  So this is the, uh, this is the, the researcher ideas, uh, agent, right? I mean, I think that's really good because again, this is the mindset you need to be building. When, when you start becoming an AI builder is, you know, how do I get maximum efficiency? How do I get maximum productivity? Well, this is another employee we can spin up, right? So while our Codex agents are building out roadmaps, marketing plans, debugging, fixing our next feature. Let's get an agent going, an idea browser that's going to now take our PRD. We can go in here, below is a PRD for my wellness app.

• 39:55Speaker 01

  Please research how this would do on mobile or how we can improve this idea.

• 40:06Speaker 01

  And I paste that in and I can hit send here. And now, right, this is the kind of AI agent employee mindset. I have my three to four employees working for me in Codex. I now have my employee working for me, an idea browser to research this idea and expand upon it. You know, maybe now I go back into chat GPT and I spin up, you know, I give another task. Hey, can you build out the landing page for the app? Right? I can go in here. Hey, please mock up for me a landing page and all the copy for it, for this app. I hit send on that and now, you know, this is how you build a one person billion dollar business.

• 40:50Speaker 01

  Is it's one person, but it's eight employees working for you at the exact same time. Our chat GPT agent, our idea browser agent, developing our ideas for us. And then our Codex agents, right, writing PRDs, writing code and building out our roadmap for our app. This is like the mindset you need to have to build a successful one person business. So this is really cool. You can edit your research brief, generate the research report. This is really cool. Okay, so it gets the challenges, the moats, the observations. This is cool.

• 41:24Greg Isenberg?

  So if you click generate report, so it's generating now, but if you, what you'll see, if you scroll down, you'll see an example of a report. It basically is the idea browser format, but for your idea, which is really cool. And if you, if you scroll up and you see at the top right, there's a build this idea. This is a new feature that basically gives you optimized prompts for your idea and for the database of ideas, you know, for marketing, email sequences, ads, all of that sort of stuff. And then you can just copy the prompts and put it in V zero cursor, Claude, whatever.

• 42:05Speaker 01

  You know what'd be really cool?

• 42:06Greg Isenberg?

  What?

• 42:08Speaker 01

  Using the V zero API and like when they generate this report, it automatically builds like just the V one, like kind of the shell of the page to kind of give that visual inspiration, like right in the, right in the report, you boom, you have your V one, copy this code to V zero.

• 42:25Greg Isenberg?

  Totally.

• 42:26Speaker 01

  That'd be sick. That'd be really sick. It's an idea.

• 42:31Speaker 01

  This is sick. Ready to build, choose your, copy your prompt, paste it in any of these builders. And oh, and this is the prompt.

• 42:37Greg Isenberg?

  Yeah, that's the prompt.

• 42:38Speaker 01

  That's sick.

• 42:39Greg Isenberg?

  And if you, and if you see on the left hand side, like, uh, marketing product, you can see what you can actually build.

• 42:48Speaker 01

  Wow, that's amazing. Sales funnel, lead map, what would the email sequence look like that we build here? This is, this is where I stink. I have 40,000 people on my email newsletter, but like I just, I don't understand the email funnel thing. Like I'm like,

• 43:02Greg Isenberg?

  Yeah.

• 43:03Speaker 01

  A dog chasing a car. I want to know what to do with it. You know, it's like, this would, this would be helpful. This is cool. Nice.

• 43:12Speaker 01

  All right, cool. And what happened? What happens when I click the tool? Oh, it just takes me to the... All right, got it, got it. Okay. Yeah, this is awesome.

• 43:19Greg Isenberg?

  Yep.

• 43:21Speaker 01

  All right, so we can now see, we have our seven different employees working for us. Let's switch back. And like, I love Claude Code, don't get me wrong. I love Cursor, but like, this isn't possible with any other tool at the moment.

• 43:34Greg Isenberg?

  Yeah.

• 43:34Speaker 01

  This is like, a really big advantage. I'm going to, uh, switch back to Codex. Let's do this. Let's see what we got here. Okay, so it fixed the error, and then let's see what else we got. Set up the AI chat. How's our agent doing here? Okay, it's still working. Oh, let's apply. Apply results, skipped, close. Okay, did this put in our... Let's see if this actually put in the chatbot and see if that's ready to go. So I'm going to switch back to our app. Let's see what we got here. Habit Flow. I'm going to refresh. Progress. We probably have to ask it to build it into our menu here, the AI chatbot.

• 44:19Speaker 01

  But we can do that. Let's go back in.

• 44:22Greg Isenberg?

  We can do anything.

• 44:24Speaker 01

  Anything we want. Anything we want. Unlimited ideas. As long as you got AI agents working for you to come up with the ideas, you got unlimited ideas. And like, this is, what you can see here is like, we went from, so far, we're like 40 minutes in, we went from literally zero idea to a fully working product. In what would probably take people weeks to do, right? You know, generally, you know, studying ideas, validating ideas, then you get an engineer to build out the, the kind of V1, they have to put it live, then you develop other ideas, and you have other people come up with roadmaps and marketing. We got all that done in like, half an hour, which is pretty incredible. Let's make sure we add the chat section to the top menu bar.

• 45:09Speaker 01

  And hit send on that, and our cloud agent will go do that. And then we can do things like, just to show the power of what we got today with agents, I'm going to go into Codex. On my browser. So say we're on the go. Maybe we're working from a coffee shop or something, or on a library computer. People are still going to libraries and we have to log, we can log into chatgpt.com/ codex because we don't have the ID on our computer. We can just go into the browser, go into codex. We can make sure we are in our wellness app and we can say, uh,

• 45:47Speaker 01

  build out the email marketing sequence for this app. And now I'm spinning up agents from the browser while I was spinning up agents from the IDE and also doing it from my mobile phone. And like, this is like the really cool way to think about AI development now. So kinda as you've seen so far, we have a bunch of different AI agents working for us. We have an MVP good to go that people can start testing out now. Now, the number one question I get other than how do I come up with ideas is how do I get people to start using the app? Well, the best way to do this is the freeway, which is just organic marketing. So if I go into Twitter,

• 46:28Speaker 01

  which for me is like, I think the quickest way to just get people onto the app, you know, you get engagement in literally seconds. I'm going to pull open Twitter, or X if Elon's watching, please don't shadow ban me. And I'm going to go post and I'm going to say, been super obsessed with wellness lately, but can't think of any habits to implement. Anyone else have this challenge? Thinking of building an app around this, right? I'd send this out. This is literally

• 47:09Speaker 01

  how I launched Creator Buddy, my social media app. I was like, I wish I had an AI that knew all my tweets. And so I went into a tweet. I said, wish I had an AI that knew all my tweets, but anyone use like an AI assistant that's aware of all your tweets and can coach you? I hit send. It went like super viral, got like 200,000 impressions. And like, that's how I kicked off my beta. So I would post this. I would then DM anyone who replies, "Yeah, I'm interested." And I'd say, "Hey, do you want to beta test this?" And I would send them the link to the app we just built. And like, this is how you start getting users and build kind of like your base for your app is, you use organic marketing with Twitter, you leverage your community,

• 47:51Speaker 01

  you reach out to them, you do the kind of unscalable, which is DMing a whole bunch of people saying, "Here's the app, do you want to test it out?" And like, that's how you get your beta testers and that's how you start building the scale, uh, which will eventually be a popular app. Like, this is how you start getting those users, is just start talking about your app on social media.

• 48:13Greg Isenberg?

  Alex, what do you say to people who are like, "Well, easy for you to say, you've got thousands of followers. I've got 500 followers."

• 48:22Speaker 01

  Uh, well, my question is this, how do you think I got thousands of followers? By tweeting, right? Like, I wasn't born and then God handed me 300,000 followers. I've been on Twitter for like, two years now. I started tweeting, people liked what I tweeted, people were interested in the projects I was building. Then I got to that level. So, how do you think I got 300,000 followers doing exactly what I'm showing you here? Which is coming up with interesting ideas and then tweeting them out. So, even if you have 100 followers, tweet out your ideas, you'll still get a couple hits, you'll still get a couple people to use your app. And like, this is how you'll get to that level where you have 300,000 followers. Like, yes, is it easier for me to launch apps because I have a huge audience? Yes.

• 49:06Speaker 01

  But also, how do you think I got a huge audience? By doing this, what I'm showing you here. So, like, you have to start from somewhere, you might as well do it instead of complaining.

• 49:16Speaker 01

  So, you put it out, you get a couple bites, like, you're like fishing in the water, you send them the link to the beta, now you got people using it, right? And so, we went from, in the span of, this has been recording for 52 minutes now, right? We went from ideas, an idea browser, we had no idea what we wanted to do. We went in idea browser, we looked at some market insights, we looked at some trends, we looked at like, "Okay, what am I interested in here? Website, save energy, solar, vomiting dog treatment, I'm going to get a dog soon. This will probably be something I'm interested in once I do that," right? You found an idea that's interesting to you. You went to chat GPT, you developed that idea, you went back and forth, bounced ideas

• 49:58Speaker 01

  off chat GPT, came up with a well-developed PRD. We went from there, we went into our building agent, right? We went into Codex, which I'm loving right now. We spun up a bunch of AI employees to develop the idea for us to build the code, to start building out a marketing plan, a product plan, a roadmap. We have all those people working for us at once. And while our employees work for us and the app gets built out, I'm now going on to X and I'm finding my first beta testers. And that's how you, as a one-person business, can have the same power, the same leverage as a massive company and start getting your

• 50:39Speaker 01

  first users and first paying customers. So it's pretty powerful. You just got to learn where you can find that leverage, like with Idea Browser and Codex and ChatGPT.

• 50:50Greg Isenberg?

  Alex, thank you for your sauce and your generosity for coming on. Highly recommend people go and follow him. I'll include links for his X, his newsletter, and Creator Buddy, all in the show notes, so people can go check that out. Anything you want to leave folks with, Alex?

• 51:09Speaker 01

  Check out the YouTube as well. All I do is post tutorials, how to build cool things as I go. And, you know, this is what I would leave people with. Ninety-nine percent of people woke up today, picked up their phone, doom-scooled for fifteen hours, and then are going to go to bed, right? If you just spend even five minutes today experimenting, searching for ideas, developing ideas here and there, you do that over the span of a week, a month, two months, you're going to be so far ahead of the competition, it's going to be unbelievable. So just find a little bit of time every day to do one of the two things, you know, we went through on this video. And over time, as that compounds, you're going to have incredible products, you're going to have paying customers, you're

• 51:54Speaker 01

  going to have something. You're going to have a lot more than what you have now. So just find a few minutes to do what we did today. I promise, cool things will happen.

• 52:01Greg Isenberg?

  And send this to a friend, because it's more fun to build with friends. You'll be more, you'll be held more accountable. So-

• 52:10Greg Isenberg?

  I don't know about you, Alex, but like when I, when I do, you know, if I'm doing something like this, like I want to do it with like the homies and I want to like build something together. So I think it's a lot more fun with friends personally.

• 52:23Speaker 01

  I absolutely, you hold each other accountable. You don't want to be the one that shows up in the group chat and they're like, what'd you do today? And you're like, oh, I just doom scrolled all day. So a lot of benefits to that as well.

• 52:32Greg Isenberg?

  Thanks a lot, man. I'll catch you later.

• 52:34Speaker 01

  Of course, my man. See you soon.