| Readwise URL | https://read.readwise.io/read/01kt2ktk7xr403asj645yptsvg |
|---|---|
| Readwise ID | 01kt2ktk7xr403asj645yptsvg |
| Date | 2025-05-24 |
| Author | youtube.com |
| Category | video |
\n\nSource: https://www.youtube.com/watch?v=jpSY4MlWX50
I prompted my way to 30,000 from an app he built on hismorning commute. How? Two tools, Cursorand Bolt. I think a lot of people makecool stuff with Bolt and Cursornowadays, but they don’t monetize itbecause they need. In this video, we’llgo over the wild story of how Polus justbuilt, launched, and marketed a SaaSproduct by himself and got traction inabout 30 days. You probably have like awinning idea already that you could justprompt your way and start building. Whyshould you definitely tune into thisvideo? Well, I had Polus share with mehis exact process on screen of how hebuilds fast with AI tools, how hecreates beautiful designs, and how youcan too. And how making money andmarketing your app is as simple as acouple of tweets. All right, let’s get intoit. I’m Pat Walls and this is StarterStory.
Welcome, Polus, to the channel. Thanksfor coming on. Tell me, what’s yourstory? My name is Polus and I used AI tobuild an app that made me $30K. It’s anapp that matches startups withinfluencers, and I built and launched itSeven months ago, it had over a thousand usersat the moment, with over 350 paid users.
I used Bolt and Cursor to make it, and Iused Framer for design. I did it allmyself, without needing a team — all in mydaily commute. Nice. Cool. Before youmade 30K with your app, tell me a littlebit more about your background, yourstory as an entrepreneur. It startedwhen I was 18 and I needed to make moneyto get my way through college, and theonly job that I could get was door-to-door sales. One day, I ended upknocking on the wrong door, and someoneliterally grabbed me and held a knife tomy throat. After that, I just had enough.
I wasn’t really employable as astudent, so I dove into every sidehustle that I could. I did webdesign, copywriting, and salesfunnels. I became kind of like ageneralist. It paid enough to get my waythrough university. When I was inuniversity, I did computer gamedevelopment. But, funny story, I actuallycompletely sucked at coding. I was veryvisual, and I just couldn’t understandit. But now, AI can actually code foryou, and so the limitations that I hadbefore are kind of removed.
The heavy lifting in terms of code andkind of patch up the area where I’m notthe best at. Nice. So tell me what gotyou into building online and realizing that you could, you know, makemoney online.
So I was going through kind of a rough patch because I triedall these different business models andI wasn’t making enough money to breakfree from a 9 to 5. I got reallyfrustrated one day that none of thethings that I did really got me where Iwanted to be. So I was pursuingcreator-led services and products justto work with creators and the idea of adatabase came to my mind, and I juststarted building it out because it’ssomething I’d use myself, and I figuredwith AI I could turn this into afully-fledged SaaS if enough people wantedit. So I really just gave it a spin askind of like an experiment, and it didpretty good.
So tell me what it feltlike to finally launch and have thisvalidated idea and have it make itsfirst dollar. Tell me about that. So formonths before launch, I was buildingthis database. It was a lot of effortand I had no idea it would pay off.
Remember, I was sitting with mygirlfriend, and we were just havingdinner. I pull up my phone. I just seesale after sale. I was like, "Okay, howlong is this going to keep going for?"And that night, I got about 20 or 30different sales. So it was a very nicefeeling that just motivated me to keepgoing and pursue this further becausenow I felt like that was validationthat this could be a real product.
So, can you share your step-by-stepplaybook on how you build MVPs with AI?If I’m a complete beginner, I wouldrecommend everyone to use Perplexity.It’s what I did. I just asked Perplexitybecause it has the most up-to-dateinformation. So it has all thedocumentation, anything that you want toany tool, any library. Just tell ityou’re a complete noob and absolutelydon’t know how to do anything, and it’llgive you really detailed step-by-stepinstructions because it’s deep research. It’ll pullup a long document that you canreference and just follow along. Thenext step is, I’d pull up Bolt and justtake it one step at a time as well. So Iwould implement the first thing thatPerplexity suggests.
In the dashboard, I import the project,cloning a GitHub repository if there is one.Once I’m in there, I just keep promptingit, and if there are errors, I just throwthe error back at the AI. It fixes theerror, I go on to the next thing, and I useBolt to a point where I get a nicelooking MVP with the functionality. Andthen I take that, download the project,and put it into Cursor. Here is whereI hook up like Supabase. It writesall the scripts to connect the back end,and I get Clerk off, which makes itvery simple to plug and play theauthentication. These are all the stepsthat Perplexity can provide for you. Soyou can just ask questions along the way,and it will give you some solutions.
Replexi is a great starting point. Boltis a great way to make the MVP, and thenyou move to Cursor to dial in thedetails. So you’re 80% of the way there,and Cursor will then just get you to thefinal finish line, having usersauthenticating. You have a back end withreal data, and with a bit of promptingand some scoring, it will make it secureas well. Okay, Polus is proof that it’snever been easier or faster to build anapp in 2025.
Make it possible to go from idea toproduct in a weekend. But if you don’tunderstand your customer or what they’researching for and how they describetheir problems, then you’re justbuilding fast in the wrong direction.That’s why we’re excited to partner withSemrush for this video. It’s one of thebest tools out there for figuring outwhat people actually want. You can plugin an idea and see what people aresearching for, what’s getting traffic,and where the gaps are. It’s got 26billion plus keywords across 140countries. And yeah, most people thinkthat it’s just an SEO tool, but trust
me, it’s way more than that. You can useit to validate your idea before youbuild, find high intent keywords yourcompetitors are ignoring, or even spinup landing pages that match what peopleare already looking for. I’m going toshow you how you can find successfulbusiness ideas just like Polus is inSemrush with one search right now. Go toSemrush and find their topic researchtool. Search any topic or niche. Let’sjust do TikTok influencer since that’skind of similar to Polus’ business, butyou could literally just search.
Anything. And I’m going to open some ofthe cards that look interesting, likethis one: TikTok influencer marketing.On here, you can see all the questionspeople are literally plugging intosearch engines. These are the realproblems people have and are activelylooking for a solution. I like this onebecause it’s not a super high keyworddifficulty, and there’s a high volume ofpeople searching for it. So maybethere’s some low-hanging fruit in there.And look right here: TikTok influencersearch platform. What’s the best way tohire TikTok influencers? Polus’business can literally be tied back to asingle page in Seamrush. And that’s justcrazy powerful if you ask me. Right now,they’re offering a free 7-day trial totry out all 55 tools in the platform.Just head to the first link in thedescription if you want to try it out.
All right, let’s get back to it. Wetalked about the code building an MVP.Let’s talk about one thing that I thinkis really cool about your stuff: Howgood it looks. How do you use AI todesign amazing-looking websites? It’sreally the simplest it can be. First,the first thing I do is I pull up Framer and IPick a free template. So any templatethat you see on there that you like.
It’s simple enough and clean for you toedit. And then you just have to pick anice font and a nice button. And thenthe rest just needs to be simple andclean. The most important part is justto get out. Focus on just editing thetemplate really quick and just put itout there. Cool.
Based on what you built and what you’ve learned working on lotsof marketing stuff and also on your own,what do you think is the mostimportant thing to focus on when you’recreating a good landing page and a gooddesign? So, I focus 80% of my time oneverything that’s above the fold. We’dcall this the hero section. This is thefirst thing that people see when theyopen the website and it has to have yourfull message just out there, andpeople really need to grab anunderstanding of what it is within thefirst couple of seconds. Anything that youcan save them time with, save themmoney, or anything that solves a painfulproblem, you should have in a headlineand a subheading, and then just a bigcall to action.
All right, so we talkedAwesome, but what about UI design? Whenyou’re building an app, how do you useAI to design UI and UX? When I’mbuilding it out in Cursor, you get thebasic thing done and then you add somefeatures to it. So, the more features youadd, the more things you work on, it’sgot to stay clean, and it’s got to staycohesive, and you got to make sense. Sofor everything UI related, I just ask itto make it clean, functional, andmodern. And I try to have it use likeShad CN components, and they’re all free.
But when I have ideas, and when you haveideas for how you want onboarding flowto be or some other elements of the app,what I would suggest is just creatingsomething in Figma. You can doodle ifyou want. You just plug it into Cursorand just explain the behavior that youwant, and then it’ll spit out a nicelooking page for you. All right. Well,thanks for sharing that. Let’s dive intothe marketing side of things because itseems like X Twitter has really beenworking well for you. Tell me about thatstrategy. Show me a tweet that reallytook off for you, and let’s talk
about that. So, this post did the mostin terms of numbers. It got close to500k impressions. And I think the reasonit worked is because I found an angle toattach it to trends happening in thespace. A lot of people, when they’rebuilding in public, never go viralbecause they never join the biggerconversation, and human attention span islimited. So it’s 100 times easier tojust bring your ideas to where the focusis already happening. So this post hada nice, simple visual demo that hasreal, raw human-like reactions.
But most importantly, it tapped into themain debate at the time, which was: is AIcoding actually good enough to build afull SaaS? Thanks for sharing that. I’dlove it if you just shared a little bitmore about the bigger strategy here, thebigger thing, which is building in public. Iwould love it if you talked about that froma strategic perspective orwhat you love about building in publicand why it’s worked for you.
So for me, what’s worked is just shipping often.Don’t just have like one big videorelease like a lot of VC startups. So,for example, I tapped into theinfluencer conversation that washappening at the time. I saw Mark Louposting a similar format.
I’d get some inspiration and thenconvert it into a post for CreatorHunter. What’s not worked for me andwhat I see is there’s been a lot ofcompetitors and copycats that come up,but they never really take off becauseeither they’re copying the exact samething, and people see right through that.You need an edge that no one caneasily copy with AI tools nowadays.Nice. That’s a good strategy. Okay, sonow the exciting part. Let’s talk abouttech stack. What tools, languages, orwhatever else you use to build thisbusiness? And if you don’t mind measking, what are generally the coststo run this business? What kind ofmargins are we looking at? I useperplexity for the game plan, bold forthe functional MVP, cursor for making itproduction ready. I use Versal, which isfree, and then Superbase also free, andthen API for scraping. In terms of costsand margins, it’s about 90%.
All right. So what I want to ask you is, youknow, you built a bunch of stuff andyou’re finally seeing success with CreatorHunter. What have you learned duringthat process that surprised you? I’mgetting a lot of questions as well aboutHow I did it, and people want the sauceon how to build stuff and build stufffast. I think you can just startprompting, and just prompt your way tomaking something and selling it. I thinkpeople underestimate the knowledge theyalready have. So, like, if you have adeep domain knowledge, you probably havetons of ideas already on what tobuild, and you probably have awinning idea already that you could justprompt your way to and start building. Ilike that. Cool. All right. Now that youare working on this full-time, tell mewhat a day in the life looks like foryou for now building Creator Hunter.
Yeah, so it’s a big change. Moretraveling and more work. When I wasbuilding out the MVP and was at the sametime traveling to China, New York,France, and seeing revenue while you’reout and about exploring is a nicefeeling. It’s a real blessing. But I’vebeen working on it full-time on CreatorHunter since the 24th of December when Iquit the job. So some days there’s morehours that I put into this than my 9 to 5.But I don’t feel exhausted at all. Cool.I like that. Well, the last questionthat we ask anybody that comes onStarter story is if you could giveadvice to young Polus maybe a coupleyears ago. What would you tell them? Ithink never underestimate how much youcan do solo. AI as your CTO is like asuperpower. I don’t know how long thisperiod of time will last, but we need totake advantage of it. Whatever ideas youhave, you can probably execute it in a
weekend. If you just figure out the mostscrappy idea possible in the mostscrappy way possible, you can get it outthere. And then if you put a buy buttonon it, share it on Twitter in the mostscrappy way or TikTok, I think you’llbe surprised by the amount of peoplethat actually want to get it. Well,that’s great advice. Thanks for comingon, Polus. I love what you built. You’regoing to be a massive success. Keepbuilding, keep going, and hopefullywe’ll see you back on the channel. Thankyou. Thanks, Polus, for coming on. Ithink his story is amazing and it’sproof that anybody right now can buildanything with AI. I mean, he just gotstarted. He built a bunch of stuff andthroughout that process found the ideathat would change everything. I see alot of people waiting for the right time.
To start, just sitting on thesidelines while others like Polus arejust getting started, playing around withthese tools, throwing them at the wall,and shipping stuff every week. If yougot started right now and startedplaying with these tools and shippingstuff, imagine where you’d be in twoweeks. Yes, it’s early and noteverything is perfect, but the earlybird gets the bag. So, if you’re seriousabout shipping and want to build somethingreal and put it out into the world,check out Starter Story Build. I put thefirst link in the description. Let’sbuild. Thanks for watching. I’ll see youguys in the next one. Peace.