Discover 3 Simple AI Tools to Start Creating an AI App Prototype

Full video transcript

Darby Rollins here with Gen AI University, and in today’s video I’m going to be sharing with you how to build an AI app MVP, minimum viable product, or at least a prototype of that product, inside of Lovable. This is Lovable, it’s the app that helps you take an idea and turn it into an app in seconds. Lovable is your superhuman full-stack engineer. Lovable has direct competition from a few other tools; whatever AI MVP tool you use, I’m going to share using Lovable to take an idea for a new app and turn it into something that has a frontend user interface and some of the key components that you would need in order to take an idea and turn it into an app. And then, of course, I would encourage you to go and try this tool out, experiment, and see what ideas you can turn into reality with a tool such as Lovable.

All right, let’s go ahead and start with Lovable. First you’re going to want to sign in or sign up with an account. Now what we can see, logged in, I have a few of my past projects saved at the bottom here in Lovable, and we’re going to start something entirely from scratch. We can show you how you would go about creating an interface for your app idea, but first we need an idea to start with. And so in order to get that initial idea, I’m going to start with a simple game of SideHustle. All right, so what we see here is SideHustle - this was not built on Lovable, but the concept and the constraints behind this game is going to help us with coming up with an idea that we can then take and visualize the prototype using Lovable. The premise of this game is simple: you pull a side hustle card and an industry card and from that formulate an idea. I’m just going to use AI to help me with coming up with that idea and finding a problem worth solving in the first place. So you can try it out if you want to break from AI and just want to hang out with some friends and have some fun.

I’m going to go ahead and randomly select a side hustle card: dig site - okay, so that’s my business name. And now I’m going to have my industry card: tech and apps. Dig site because we’re going to be building an app - dig site, tech app. So now I can come up with the idea. It could be a lot of things - I see like a way for a… to visualize maybe a site that I want to dig at to find buried treasure or something along those lines.

Before I even do that, I’m going to go on over to Perplexity. Here inside of Perplexity I have the ability to do deep research on any sort of idea or topic. I’m going to say I have a business idea called Dig Site in the tech apps space. I’m going to go ahead and have Perplexity do a little deep research, because I don’t really dig for treasure that much, but what I do know is that there’s a lot of people out there in the world that do dig for treasure or do look for treasure - it’s a hobby - and so maybe there’s some validity behind this idea, and maybe there’s an app that could make it really easy to plug in a few pieces of information about maybe where you’re at, where you want to dig for treasure, and maybe the AI can search and give you an idea of where you can go dig. So let’s see what Perplexity comes up with here - I like watching Perplexity do this because to me it’s cool to see the thought process going into it. It’s found several innovative approaches to AI and treasure detection, including smartphone connectivity, augmented reality features that enhance user experience - however, I still need to investigate. And so this was just again an initial idea from SideHustle: dig site and tech apps, the concept at a high level. Right now I want to do a little research very quickly and see how we might structure an app like this that we could build with Lovable that might actually have an audience that has a pain point that we can build a solution for.

Look at this - over here is “AI: The Future of Metal Detecting,” but here we see “AI-Powered Treasure Hunting: Building the Dig Site App Business.” The convergence of artificial intelligence, geospatial technology, and mobile computing has created unprecedented opportunities for treasure hunting enthusiasts. We might have something here, y’all. So you see the Dig Site mobile app would serve as a primary interface between users - location-based treasure hunting experience could form another cornerstone. The app could generate personalized treasure hunts based on the user’s location, guiding them through a series of clues and predictions and discovering sites - “Building the Gen Treasure Hunt App for Curious Explorers.”

So now I’m going to say, based on this information, what would be the highest impact, lowest lift core - one, two, three - features that would be most applicable and useful for people who are interested in this type of app to help them with their treasure hunting. I’m giving it a little bit more of a prompt to lead it onto the next step, which is: based on all this information, what do you think would be the highest impact thing that requires the lowest lift, one to three core features that we would find.

While we’re waiting for Perplexity, I’d like to draw your attention to NotebookLM by Google - really helpful tool for those of you that learn best by audio, hearing podcasts, and listening to people talking through concepts that you’re learning or want to dive deeper into. You can simply copy and paste or link relevant sources like Google Docs, folders, YouTube videos. I just copied and pasted that initial research in here and my folder now has an AI podcast being generated - it’s probably going to be anywhere from seven to ten minutes, maybe less, maybe more, I’m not sure. It’s generating this conversation for me to listen to, to make a podcast about it, if I really wanted to digest this information and develop a deeper understanding of this topic. And so that’s just happening in the background now.

And as we can see, deep research searching the web found several sources discussing the importance of interactive and gamified features. Now we see core features for Dig Site and AI-powered treasure hunting app, and again we’re going to tie all this back to Lovable, but what’s the point in going and building an app if we don’t even have an idea about what problem it’s going to be solving? That conversation still generating, and here we are with the core features. This report identifies three of the highest impact, lowest lift core features: one, location-based treasure probability mapping foundation - should be a sophisticated yet intuitive location-based treasure probability system. This feature could leverage GPS functionality to identify high-potential areas for treasure hunting in the user’s vicinity. The system would display a heat map overlay - so that would be interesting - highlighting locations of increased likelihood of valuable finds based on various data points. So there’s a location-based system, there’s a real-time visualization interface, there’s a community data contribution system to continuously improve these treasure hunting experiences through the power of crowdsourcing data at large scale. Some good ideas here: three core features - location-based probability mapping, real-time visualization, and community data contribution - that would be something I would probably want to partner with a community of some sort that would be a part of this as a user base.

So I don’t know if that’s going to be the first feature - I’m wondering, if we had to cut anything out of here, if you were to assess just one of these features to solve an immediate problem for people based on research for people getting into treasure hunting who are more casual about it, what would the top pain points be? We would solve painful problems that we can create one feature to solve in a simple, streamlined way. Deep research is going back again to narrow the focus of what we’re going to be building here, and my twelve-minute-and-twenty-five-second podcast is now ready to listen to. So yeah, I got my podcast - I can listen to treasure hunting while I’m building an app about treasure hunting.

And once this next bit of research is in, I’m now going to take it and create a project specifically for this app inside of Lovable. Honestly, you can go and make a bunch of half-baked apps inside of any number of these tools. I wanted to demonstrate the research going into it first, because again, what’s the point in just making one of these apps if it’s not going to even remotely solve a problem? And so when you’re going into a tool like Lovable, you’ve got an idea for a specific problem that you’re already looking to solve for, or you could just vibe code and play around and have some fun and see where it takes you - nothing wrong with that. And this, to a degree, is going to be my version of vibe coding, which starts with the research.

Primary pain point for casual treasure hunters: location intelligence. After analyzing the research of treasure hunting, one critical pain point emerges for casual enthusiasts that can be immediately addressed. The research clearly reveals that one of the most frustrating aspects of treasure hunting for beginners is wasting time and resources on unproductive locations. According to experienced treasure hunter Doran Cook, only about 3% of all lost treasure stories are worth giving any consideration or hunting them - this creates an immediate barrier to entry and enjoyment for casual treasure hunters who lack the experience to distinguish promising locations from dead ends. Many novice treasure hunters spend excessive time, energy, and resources exploring ideas or areas with minimal potential, as noted in this treasure hunting guide: if you can’t dig deeper in researching a story than all the others who hunted it and uncover the truth about it, hunting it would be an exercise in futility. This leads to disappointment, wasted efforts, and ultimately abandonment of the hobby before experiencing the thrill of discovery.

Location intelligence feature solutions: so the most impactful feature to address this pain point would be a location-based probability mapping system that helps users identify promising areas near them using AI to analyze multiple data sources. This feature would serve as a digital equivalent of the experienced treasure hunter’s intuition. The system would analyze several key data points: historical information about the areas, including former settlements, trade routes, known historical activities, geographical features that would typically correlate. I feel like AI could do a great job at using tools like Perplexity, OpenAI research, and otherwise for the research side. Here’s the features - the features directly address the insights that if you are not yet adept at the detective work, this will better improve your clue-finding efforts. Essentially automating much of the detective work - this is going to be like a location detective app, right, Dig Site - let’s do some detective work to help you with that experienced treasure hunter intuition that can help you with your treasure hunting adventures. So that solution is technically feasible with current mapping and AI technologies available in 2025 - it builds upon established practices of the treasure hunting community while making them accessible to beginners. Most importantly, it solves the most immediate problem preventing casual hunters from experiencing early success: knowing where to look in the first place.

I like this. Now what I’m going to do - I have Lovable here, I can do two things: I could take this concept and feed it to Lovable and let its AI work through and help me with processing it and helping me visualize and create that, or I could use something like Claude to help me with coming up with the initial prompts and seeds that I’m going to provide to a tool like Lovable. So I’ve got the research, I’m going to create a new project in Claude. Create the project, and I’m going to add text context, give it a title, and I’m going to paste that context in here. And I’m asking Claude, based on the context of the problem, we are going to create an MVP that solves this with a simple AI integration about their adventure, and then the AI finds relevant, vetted locations near them. Use the context from the research provided about the problem to create the app main page, website positioning, the login, and then the main app dashboard, which is centered around this one key function. I’m using Claude 3.7 Sonnet for this.

Now Claude is going to work, and it’s actually helping with some of the code here, but what I want to have it do is something with an initial mockup and fleshing out the information to then go put it into Lovable and turn this into something that’s actually functional - mocking it up here, and you can see how impressive Claude is too at helping with getting some of this stuff out and running. You see: “Only 3% of treasure hunting locations are worth your time - RAI helps you find them, start hunting now.” Not a bad play. Claude’s just going another step further here, which frankly is probably a little further than I wanted it to go, fleshing through the idea a little bit and getting a good structure for how I’m going to want to see the app flow, and then bringing it inside of Lovable.

So we can see it actually made a homepage - treasure map, recent hunts, settings, high-impact potential, your location, hunting context, tell us what you want, find the treasure location, your AI treasure hunter. See the landing page: stop wasting time on empty digs, why treasure hunters choose Dig Site, AI-powered location intelligence, successful adventures, success stories, find real treasure, create your account now.

I want you to write me out a multi-step prompt wireframe which I will use to inform building this out as an app MVP prototype in Lovable.dev - the outputs you provide should be easy for me to copy and paste to Lovable, then I will work in there for the rest. So we see the Dig Site app prototype prompt wireframe - so now we’ve got our project overview, project tagline, description, target user, pinpoint app screen structures - landing page, login/signup, dashboard, location details, settings, and user profile - and then we see the landing page copy, the login/signup screen, dashboard, main location details - I don’t need the roadmap details - visual guidelines, function requirements.

All right, so now I’m going to copy this and jump into Lovable, paste. Now let’s see what we got cooking - noting that Lovable is going to work, it’s using Tailwind, so we can see it’s going to town writing code - see if we can’t bring Dig Site to life. Yeah, I would not know how to even do any of this without writing all this stuff out. Anyways, good thing I can help code, but there’s going to be some vibe coding, going back and forth, getting it to do what we want it to do - giving it some API keys, maybe it needs a Google Maps integration, maybe we can integrate a Pickaxe tool for some of the search functionality, building out the testimonial section.

And make sure you guys are subscribed to us here on YouTube if you like this type of content, if you like learning about AI tools and their applications, whether you are looking to build your own or just looking to see what’s possible - Gen AI University is here to help with bringing you some of the latest of what’s going on, not only from myself but other experts and insights from those experts doing the things with AI inside of our community and through our Scale with AI Summit series and podcast, which can be streamed on Spotify, Apple, and everywhere else that you listen to podcasts.

The moment of truth - okay, well that’s a white screen of nothing, let’s see what it said. “We’ve created a stunning site app” - yes you did, Lovable, beautiful, well done. Refine, customize, tweak, master prompting, iterative prompts, better outcome, expand the backend - yeah, so we’ll connect with something like Supabase, GitHub sync to sync edits, debug with ease. “No, why was the build unsuccessful, fix the UX, fix it Lovable, I don’t have time for this, did not work for me, I’ve got treasure to find” - yes, see it noticed the build error right here, so it’s going to go and try and locate it. See it’s importing icons from some tools using React - okay, you replaced it, but I still don’t see it - help me out here, Lovable, we came all this way, we’ve got a potentially multi-billion dollar app on our hands, we could only just get started with the MVP. Obviously this is a demonstration, and if I were to go and actually want to build this out for real, I’d want to partner with somebody who already has an audience of treasure hunters that we could validate with.

So it did quite a bit in that one prompt that we gave it - maybe too much, maybe we broke it - got to fix things up a little bit here, troubleshoot a little bit, see what’s going on with the pages. It’s working - well, Lovable, I can’t navigate, I can’t see anything, I’m going to refresh my screen, maybe it’s a user error in this case. Whoa, look at that - wow, “Stop wasting time on empty dig sites, only 3% of treasure hunting locations are worth your time, RAI helps you find them, start hunting now” - this could use a little bit of work. AI-powered heat map showing potential dig sites - I don’t know how we’re going to build that. “Treasure hunting reimagined, stop relying on luck, use data-driven insights to find historical treasures, AI-powered intelligence, start hunting now.”

CTA features, dashboard - I’m in the dashboard, I have treasure map, recent hunts - that page hasn’t been built yet, home - interesting, I don’t know how I would build this - high-potential dig maps and location and areas, back to the dashboard - this is basically what Claude’s interface looked like too, pretty much replicated it. Your AI treasure location finder, find high-potential locations, our AI historical data - this stuff’s going to be like the backend, like your location - I want the dashboard to be more simple, and for the MVP we will just have users input text location and hunting context, and then we will output our recommendations based on our findings via text, no images.

Yeah, have treasure map, recent hunts - obviously there’s a lot that we could do here, but I just need one feature that’s going to be the breadwinner, so to speak. Now we’re working through our treasure map, it’s thinking through it - see, treasure map default there, keep it essential, go ahead, reset this up, dashboard, just to optimize it. No, Lovable, look how far I got with a free account - I didn’t even have a starter account right now, which again it’s only $20 a month, probably worth it - but that basically got me up to a point where I could visualize and turn this into something that I could probably build from, obviously without an audience, without a community, or without a partner that I would be able to launch this with, I could build a cool app that no one would use, and wouldn’t that be a shame.

Now with something like Lovable, you can see how fast I can get a dashboard and some stuff set up - there’s still going to need to be some work that gets done with it, but at the end of the day, maybe this could be the next big thing for treasure hunters. Stop wasting time on empty dig sites - if you’re ready to find real treasure or make your own AI app MVP, comment, let me know below, make sure you like and subscribe, and if this gave you any ideas for how you could potentially use Lovable plus other tools like Perplexity and Claude to help you get started, let me know in the comments.

That’s about as far as you can get with a free account, which to be fair is quite a bit without even giving them my credit card, and now you could upgrade, vibe code a bit more, and you could take an idea like Dig Site and connect it with a backend like Supabase, come up with your very own AI app MVP. Good luck on your next prototype, and as always, keep calm and AI on - I’ll see you on the other side, cheers y’all.