We Built An App In 30 Minutes With No Code + AI (LIVE)

Patrick Biddiscombe builds a HIIT workout generator app live using Replit's AI coding agent, which writes code and deploys it in 30 minutes. The app takes three inputs (energy level, duration, focus area), generates a custom workout plan, and outputs a downloadable PDF. Replit Core costs roughly $130 a year or $10 monthly.
Questions this video answers
September 19, 2024
$130 per year (noted as about $10/month on annual billing)
$12.80 per month
streamlit, ReportLab (ReportLab spelled in transcript as 'Report Lab'), and SMTP/email-related libraries
Energy level (low/medium/high); Time (15, 30, 45, 60 minutes); Body focus (lower, upper, both)
Chapters
On today's show, we're talking about one of the most important AI releases of the year, and it's probably flown under the radar for you. It is replit's new coding agent. We're going to use it to build an app live on the show today. While we're recording, it is taking coding from hours to minutes. It's ushering a new era of personalized software where everyone can build tools not for hundreds or thousands of people, but just for themselves because it is becoming so fast and so cheap to do it. Let's break down everything you need to do and we'll show you some real AI coding all on today's show. Kieran, we have said this was going to be the year of AI video. It's turning out to be the year of AI coding. And there is a massive wave of new AI coding tools. And I'm here to share with our audience that I think we have found something that is a better value in AI than even the $20 a month for something like ChatGPT or Claude Pro, because the team over at Replit have just launched something pretty killer. Replit is an app that lets people code and tries to. Is trying to make coding accessible for everyone. So naturally, what did they do? They released an AI agent to build web apps for you so that you can build web applications without being an expert coder. Or even if you are an expert coder, you can turn your ideas into apps much, much faster. Amjad, who's the CEO and founder over@repl.kieran had like the best way to describe this. He tweeted that we just violently entered the era of personal software. Could you maybe explain to everybody what he means by personal software? Because this, this person right here, he's showing just built a personalized workout and diet generator on replit.
Yeah, I think. And we are going to try to build something on this episode live with repl Dot Agent. So you want to stay tuned for the app that we build because what, what we've talked about in this show is a couple of things that have started to come true. The first is search, the second is video, and the third is one to one software. And so in a world where the ability to just code for yourself and build anything you want yourself is available to you, it completely changes the way that we think about software. Like, it turns software on its head because forever, we've had to basically create software for one too many because that's the way unit economics will work, right? Like, you actually have to put a lot of money into this. Then you have to be able to sell it to a lot of people, and you have to have them using the product for a long time. When you go up market, right, you can start to do real custom things because the contract value is so large. Then you can expend a bunch of capital on engineers to customize it for that company. But there is no doubt that all software, if everyone had their, you know, wishes come true, they would probably get the software, whether it's Strava, whatever kind of software, and customize it a little bit to their needs. And what we're seeing with these agents is if you have a lightweight use case that you really want to use software for, you kind of no longer have to go look for a product that does an okay job of fulfilling that need. You can just ask one of these AI agents to go, literally go and build the software for you, and you can just code with that agent through the English language. And Replit agent is another huge step along the path to being able to have this incredible world of personalized software where we're all building software to do tasks for us. And I think actually, in the future, you don't even think about this stuff as software, right? Like, you just have a personal AI assistant and you ask the AI assistant to do things for you and it codes things in the background. And you don't even think of that stuff as software. You just think of it as an assistant that can do most things that you want.
I guess I think that's the perfect articulation, Kieran. And here's another tweet that kind of, I think, kind of helps bring this to life from Kyle Shannon. He's using this new Replit tool. He's like, I've had this idea for 15 years and I built a functional version, 15 minutes. And what he built was a personal app for himself where he adds his memories to a map so that, like, all of his memories are, like, tied to, like, physical locations. And he can kind of keep track of his experiences in a map, something he's always wanted to do. And he walks through, like, a full demo of, like, how to do this and how to do this very quickly. That's something where even if you had the coding knowledge, you probably wouldn't, like, have spent the time necessary to do it, because you're like, oh, this is a cool idea, but I'm not going to spend hours to go and do it. Where if you're like, hey, I can maybe spend 10, 15, 20 minutes and do it. That's a whole different equation and that's what's interesting. You want to know what's really interesting, Kieran? Do you know how much this costs?
Well, they do a really good SaaS annual to monthly discount trade off. But if you do the annual, it's $10 a month.
Yeah, it's a. Right now it's $130 a year. If you pay for the year to get replit core, which includes the AI agent and all the other kind of core functionality of replit that is absurdly cheap. It is cheaper than your chat GPT subscription or your CLAUDE subscription. Like, again, it's not just the power, it's the cost for the power that continues to blow my mind. Like, if you had gone back five years ago to us and said, hey, you can just write sentences and build a web app, what would you pay for that? We would have said, like, take all my money. Literally, just take all my money. That sounds great. I would totally pay lots of money for that because that is a great way to learn how to code, move faster, build things, see if something's a good idea. Right. And that is the era we're in today.
Right. The cost of this stuff is coming down all the time. I do think it's interesting all these kind of agents are somewhat in competition with each other. Like, Claude does a pretty good job of coding. You've got replit agent. Replit agent does some special things. Like it can set up all the databases and do like, actually help you launch the app, which I think makes it differentiated from all the rest.
That's why it's better. It's not just like some locally hosted version of your web app. It can help you launch it. It can install the libraries you need. Does a lot of magic in the background for you. Right?
So we should actually get into build something.
Let's do it, man. I got my account here. I can pull this up. All right, so you log into replit, you get into replit Core here. And if you're a replit core customer, you get early access to this. And Kieran, why don't you be my customer? What do you want to build today, my friend?
Let us do a little bit of the Alex Lieberman, guest of the show to do. It's like one of the most common use cases is a fitness app. I want to have a hit app that generates a hit workout based upon your energy levels and the time you want to do that thing and then the area of the body that you want to. You want to use. So basically we have three core inputs. If the user inputs energy level, low, medium to high. The user inputs time. And I think we put like 15, 30 minutes, 45 minutes, 60 minutes. And then we put the user inputs lower body, upper body, or both. Energy level, how long, how much time.
And then area of the body focus on lower, upper body, upper, upper body or both. Okay. Okay, Kieran, we've given it some basic instructions in our prompt here. What are the other additions just besides the very, very basic inputs of the app? Like how do you want the output, what format do you want the output to be? How do you want it to figure this out? What other instructions should we do here? You're the customer.
Okay, let's say we wanted to output a hit exercise plan for that day dependent upon the user inputs. And we want it to be downloadable as a PDF so I can have it on my phone. And we want it to be emailable so I can share it with people that day. I think we can start with that and see how fast we can iterate on it.
Let's see what's happening here. It's telling me this is an experimental product. It's still early access. All right.
You know one thing I'll say about the replit launch that I love. Yeah, they just launched it. You know, all of these kind of AI launches is we're going to launch this thing and then it's not available for like six months. I think the best launches is just launch it, get it out, get it out there, do it.
So one of the things that I like about this, I've used it some over the weekend and I will say the other thing I like about this, Kieran, as it's, as it's coming back with some recommendations is that the mobile app is very good and what I'm trying to train myself to do is instead of like consume social media, when like I'm waiting in line for something or have like 5 minutes to kill that I actually start building something with those five minutes instead. And that is like a starting to be a life changing experience where you're like, oh my gosh, the utilization rate of my time is just way better when I'm building versus consuming.
Yeah, yeah. Having I think I mentioned to you as off mike and this is nothing to do with, this is no, no indication of how valuable X can be. Like I think there's some good things on there. But removing X from my phone was the greatest thing I've done this year in terms of like life, happiness, productivity, all of those things, I just feel so much better.
Yeah. And you know what, then you switch that time also to building something and it's like a double win.
Yeah.
And so, so with tools like this, I think one of the things we would encourage everybody to do is spend a little bit more time building than you are consuming right now. So Kieran, we gave it our prompt and it says, here, let me build a plan for you. It's one of the things I like about this agent and says, I've created a plan for a HIIT workout generator app using Streamlit. So it tells you what's technology. The app will allow users to input their preferences, generate a workout, download it as a PDF and share via email. Would you like to review the details of this plan? And we say build an initial prototype and then it gives you options for additional features it can build, add a database to store user profiles and workout history, implement a more advanced workout generation algorithm, add video demonstrations for exercises, and create a mobile app version for offline access.
Oh wow. Do them all integrate. I say, why not? Like, it's all free.
Let's see, let's see.
Imagine the developer. You know that if you were a developer and you came back, but look, you asked me to do this thing for $30,000, here's the other things I can add for no extra cost. You'd be like, oh my God, that's, that's basically what we're doing.
So this is pretty wild, right? So all right, I'm going to do that. And then there's a button says approve, plan and start. So I hit this button and I run the prompt and you can see.
It starting to code, right?
It's starting to code. It says that's building the initial prototype. You can see it writing the code right here, right?
Imagine they found out. So the founders of Stripe or Skype, I think they are from somewhere in Eastern Europe. And their second, this could also be a total made up X story. But their second or one of their next startups was this robotic company. I think I might have told you this story before where they we're gonna have robots delivering food and they had it pilot piloted in a city somewhere in Eastern Europe. And you would get photos of all of these little robots running around delivering food. And it turned out that in the back end there was no AI control in it, which was there was meant to be. There was just a bunch of people with like remote controls to test the idea. And it'd be funny if Replitz was like all of these Coders just like, they're just rapidly, like trying to code.
Dude, as fast as this is freaking. That would be freaking impossible. As you see, it's writing an amount, a large amount of code, right? And it's installed multiple things. It's installed streamlit, Report Lab, smtp, LB for email. And it's just like building away as we're talking, right? Which is pretty crazy.
This is about. This is like the. As much code as I wrote in my four years of university and it.
Literally just wrote it in like two minutes.
In two minutes. God, it's just so. Such a different world if you are a coder.
Have you used this yet? Have you. Haven't used.
I. No. I signed up last night. My weekend was not. I did not have the ability to do some code in at the weekend, but I signed up last night because I have an idea for an app. I basically want to create an app. I'm gonna try to create an agent version and an app version just to see the difference. But basically all along the same themes that I've been doing. But be able to take a YouTube transcript, divide it into talking points and then if you. It puts those talking points in the tile cards and in those tile cards you can click LinkedIn, YouTube, short or X whatever. When you click on all the tiles and you hit next and it will create content and for you in that format for that platform. So that's my thing that I'm going to try to build.
That's very cool. Okay, Kieran, it's come back with us with the prototype. You see it right here.
Okay, let's test it out.
So how's your energy today?
Low. So tired.
How long. How long do you feel like you could work out today?
I'm going to do 30 minutes. 30 minutes.
What do you want to focus.
I'm going to do upper body. Let's do a bit of upper body.
Upper body. Okay. You want to see how it works? See if it works?
Yes.
Here you go. Here's your workout, Karen. Inchworms.
Can you download it?
Yeah, let's see if you want to see if it downloads. Yeah, it does.
Darn. Here we go. Personalized software. There you go. My own hit app. I just have to actually go run that on a.
Here, let me, let me, let me, let me. Let me try to show you the PDF real quick. So does it.
It deploys it so you can share it as well. Right. Like they host the app, very similar to Claude. So you can basically just start using it from day one.
I'm going to show you in a second here. So window we need.
You know, way back in the day I tried to build that app. I had a domain name. I had like kind of had 500 odd domain names in my time, but I had a domain name for it. I coaxed a friend to like help me build it and the bloody agent has buil it in five minutes and I never got it done.
So this is, do you see the PDF output here? This is your PDF. So it's very simple PDF chart that you could share email. Right.
Plank downward dog Plank Jacks jumping jacks brawls. Okay. Rest I would say. The only thing is I would iterate with it because I don't like anything that has 22nd work and 42nd rest. That's for, that's for, for people who just are not serious about working out.
Well, Kieran, let's go back to our app for a second. Maybe it's smarter than you and maybe if you have high.
Oh, my energy level. I thought I did lower energy level. I thought I was just. There you go. I'm sure.
Look. And that's the variable it changes. Kieran.
Ah, there you go. High energy.
45 seconds of burpees, 15 seconds of rest.
Okay, I'm gonna do high energy. I'm not doing loser stuff. 40 seconds of rest.
What are we doing? And then it does. And then it has a little blank in here that you can just put an email address in and share it. And you see up here you can deploy your app.
I assume the email doesn't work right.
Now, but let's, let's look, let's see if it. I can put it to my personal email that hopefully the producers will block out so I don't get a bunch.
Of emails, cell caps and stuff with your virtual AI reps.
I'll tell you, I'll tell you in a second if it comes through.
Okay, real quick. On this show, we help you save time by pinpointing the AI strategies that actually deliver so you can ditch the ones that don't. To help you stay focused, we just dropped the new AI Trends for Marketers report. This resource is packed with original research and findings on the latest trends in the AI space. And we've got some really surprising insights that you can only find right here. Now you're going to learn the top five AI tools across all of marketers. Who wouldn't want to know that information? We're going to tell you exactly how teams are investing in AI and where they are seeing a high return on Investment. Is AI going to help or hurt your marketing career so you can steal hours of research for free? Just grab your copy from the link in the description below. Now we're going to get back to the show. All right. Why aren't any of us at work anymore? Should we just quit our jobs to do this stuff? What's the point of doing. I don't even want to. Why am I showing up to work every day? Like, there's agents out here building apps in 10 seconds.
Like.
I'm just going to go enjoy life until these things just start running everything for us.
Coding. So we want to see what happens if we hit deploy.
Yeah, let's deploy it so we can share it. Because I want to use. I actually do want to use this app.
Okay.
I can actually iterate.
We can deploy to production.
Can you share code? Can you actually share the entire thing with me and then I can just start to iterate and build on it or they don't have that feature.
No, I think I could share the code.
Dope.
I know I can. I do it bare minimum. Yeah, I think. I think we can be teammates, too.
There you go.
Where we can build and. Okay. Do you want to just do what they recommend for a reserved vm?
Yep.
I don't know how to do. I don't know how to do this.
I. Can you just click next?
Yeah, it's $12.80 per month. Sure.
Oh, okay. So you need to pay hosting fees. That's pretty clever of them, actually. That's pretty cool.
You got to pay some hosting fees. This is how my. I'm going to spend hundreds of dollars a month building web apps for myself. For your life.
The one to. Yeah, I want to.
Other.
I need to start to think about, like, other apps that I would just never pay someone to build because they would just be so little usage that I actually want.
I feel like this is like a whole separate show. Kieran, Is that we should each come with like three app ideas that we've built and show them off to kind of help people understand what's actually possible. Because it's pretty wild.
Yeah. I've been bookmarking people stuff like crazy to see what other people will have been building. Everyone has built. Literally everyone has built a something to do with stock trading app. I see that one as the most common. But I want to show you. Actually, let me just show you a couple because there's a couple that are like pretty running. So there. Because there's a couple here that actually would make sense. There's a road trip planner. Let me just show this.
Oh, I like that.
There's a road trip planner. So this one, this person built an app and you can put in where you're going and it actually will plan out a road trip for you. Well, this is a big thing in the States. Travel and actually like where you should.
Stop, what activities you should do, trips.
Yeah, like all, all of that kind of stuff. Again, that's just like an example of a personalized app that you would potentially never have someone build for you, or you would not have someone build for you, but you would want to customize in your own way. One of the big ones I think is going to be your ability just to build web pages. So this person built out like an entire. A whole webpage waiting list. Everything you can use from day one for a startup to start to actually collect email addresses. And it's just so quick at doing it. Like, I don't. I expect there's so many small companies are going to use this to be able to just code out their own things just so faster than trying to get outsourced to get an engineer to do it. So for people on the rss, we're showing someone live building a land page and it took them under four minutes from start to finish to have a startup landing page with all of the things they need plus a waiting list to actually start to subscribe to their software. Let's see if there's any other good ones. Okay. Here's one of my favorites, actually. Someone I actually, I thought of building this.
No, you always love your favorite throttle is like, I had this idea, it's.
Something I would never build because I just don't think it would like, it would have been, you know, something that would have been worth investing in to try to build a business around. But it's basically a version of Stack Rank for prompts. Right. You can upload prompts. And so, like you can see here that this person has built an entire website.
My friend Martin. Shout Out Martin.
You know him?
Yeah. And he's, he's. He's cool dude. He's an og. Okay. Social. Social guy.
There you go. So he is building an entire kind of Python. No, an entire kind of prompt website where you can kind of up, up, vote and submit prompts. I think that one is pretty cool. Like a prompt library.
All right. The other thing, Kieran, that you can do that I, that we had talked about off screen, but I had just decided to do this is you could. You can give the replit agent. I said, hey, what app should I build with the similar web API? And I just sent the link to the API documentation and I said, can you give me four ideas? Let's see what you think of its ideas. Karen, I'll propose what to build for you. I propose four innovative app ideas using the similar web API. Would you like me to present them to you? Oh, yeah. It offered Karen to present a detailed market analysis for each idea, user Persona development for target audiences, mockups or wireframes, and preliminary API integration plan. Yes, to all of these.
Right. But if it's going to do one, just do the second one, the Persona building. I want to see how it does that.
All right, let's do it. Ready?
Also, I just realized on the one that I built, it didn't integrate the features it said it would integrate. Didn't do a mobile app, didn't do. What was the advanced stuff that it said it would do?
Oh, yeah, I don't know. I didn't know how. We actually. We actually didn't look at it enough to like.
It's like a typical engineer promises you things and doesn't even do them. So it actually is really like working with a developer.
Man. The agent encountered an error, Karen.
Oh, my God.
This is what you get with software.
Well, you know the funny thing, the funny thing is you know how humans reset their expectations. So basically to your point, like, if someone told you you could get an entire app coded in 10 minutes a year ago or even like three months ago, you would have been, that is unbelievable. That is magic. I just want to live in that world. And then you live in that world and then humans are like, yeah, it's got an error. Rubbish.
But Kieran, do you want to. Do you want to see something?
Yeah, let's do it.
Look, you know what I got to show you. I have to show you our app live on the Internet.
Oh, there you go. Just need to put it up. I bet you could add a payment screen as well.
Oh, for sure. So look.
Very dope.
It literally in the time it took us to do the pod, we built an app and deployed it on the Internet and literally use it. Just go and use it now.
We'll put it in the link while. Link in the. In the YouTube so you can go play around with it.
But isn't that like, how unbelievably fast was that? And how like, mind bending is it that it was that quick?
This is. I'm telling. This is like, again, this is bring in the personalized software to Life. If you don't see the future here, everything gets personalized to one to one software. You can build one to one software media. You can. We talked about this in the last AI episode where you can. When the video tools get good enough, you can just ask Claude to write you a script or a version of a movie you like, write it in a different way and then plug it into the text video tool and have your own movie. We were already seeing one to one music. We're going to see one to one. All forms of media are going to be one to one. And this is a example of one to one coding.
This is personalized software for just a fun personal idea. But like looking at the similar web thing, it's like if you're a marketer, there's a bunch of marketing data similar web has. You could say, oh, for my company, I just want to build a couple of apps that are just data around my company that I just want to like track and understand or my company versus a couple competitors. All of those things. You can now do that very quickly, very low cost.
Yeah, it's great for internal tools as well. Like trying to build like all of these internal tools that you could build for your company again, that would. You would never have engineers to be able to do it, developers to be able to do really helps you build, you know, those kind of tools.
I have seen a lot of people doing it for internal tools and building them faster. Okay, so Kieran, back to the similar web idea API idea. It had some issues, so I said just give me the ideas. And here are the four ideas. Competitive Intelligence Dashboard, Industry Trend Analyzer, Influencer Discovery Platform. SEO Opportunity Finder.
Great ideas. Better than most of us. I think given it APIs, I'm actually going to give it the Glassdoor API because I was playing around with some ideas there to like tell me what apps it would build off Glassdoor data or I'm going to give it a couple APIs. I'm going to give it Glassdoor similar web and I'm going to give it maybe even G2 and figure out how I could build a pretty incredible company dashboard.
And then I just asked it, hey, can you build number three, the influencer discovery platform? And let's see what it says.
Let's see what it says. I want to end on one thing for if I can just share my screen. If you are a coder and you're like, oh, these, you know, this episode is about our role and us being heavily disrupted. I think that this just pushes developers up the stack a little bit. Like maybe lightweight tools become easier. I think there's still obviously a need for developers and actually it makes people with deep domain expertise just much, much better. So developers can truly become 10x, 100x versions of themselves. I think this is a 100%, this is a great tweet to end on that. Like, there's still a big gap between technical and non technical folks. Where you had this person post this interaction with his friend and the friend is like, oh, I just built something. This is amazing. I'm now an engineer. And they're like, hey, share it with me. And then he tries to share the. Or they try to share the local host, which is just the local host version of that on their laptop. So there's still like a big gap. If you're looking on the YouTube, you'll get the joke a little better than RSS. But if you're, there's still a gap between all of, you know, a big gap between the people with deep expertise and the people who are beginners. But I think it helps is the beginners to actually start to do things for themselves to a certain point. And it helps the people with deep domain expertise, 10x their productivity, 10x their output. And I think that's pretty awesome.
I think that is the right insight is that it really helps both ends of the spectrum. Right. Because when you're more on the novice on something, the first and most important thing is just like understanding enough to know what's possible. And these coding agents really help you do that. But if you are an expert at something, wow. I can imagine if you're just like, hey, I got this idea and I can just have a first run at code. The other thing, Kieran, that I have seen a lot of people doing is just finding code that isn't working and telling it to fix it.
Right, Exactly. Debugging, like, hey, I got this code.
It'S not working for some reason. Can you figure out the problem and fix it for me? Right. So even if you are like a very seasoned developer, you can just bug fix and fix problems much, much faster, which is much easier.
Yeah, you can, you can have the AI assistant do all the things you likely don't want to do, so you can actually spend more time creating things.
Yeah. But I just have to say, Kieran, what a great freaking time to be alive, my friend.
Pretty liberating that you can go from idea to fully fledged thing pretty rapidly.
It's pretty incredible, right? And like this is pretty awesome. Kieran, anything else you want to add before we close off today's show?
No, I'm getting prepped to do my hiit workout, so I've got a good hiit workout. Gonna have a good evening for myself.
We'll put the workout app down the link in the show so everybody who wants to work out with us can. I'm gonna be using it too. I love a good hiit workout and it's been a fun show today. Anytime we get to, like, use AI live on the show, it's always a lot of fun. And we'll see you real soon. On Marketing gets the green. This data is wrong every freaking time. Have you heard of HubSpot? HubSpot is a CRM platform where everything is fully integrated.
Whoa.
I can see the client's whole history. Calls, support, tickets, emails.
And here's a task from three days ago. I totally missed HubSpot.
Grow better.
More from Patrick

A NEW BREED OF PARTNER: The Partnership Ecosystem Model Journey of Patrick Biddiscombe, Part 1

Customer Success Workspace: Never Lose Sight Of What Matters Most | HubSpot Spotlight








