Ever since AI started coding well, a lot of people who had never coded before started building their own products. People started building apps that solved problems they were facing, which they couldn't do previously because they lacked skills that were only limited to developers. But these weren't just hobbyist side projects. They turned into serious products and a lot of them started pulling in real revenue, not just in thousands, but in millions of dollars. This all was able to happen because AI bridged the gap that was there before. But none of them got there just like that. They all followed a series of steps to make it work. They didn't use some workflow that nobody else can use. None of them had developer experience or had business experience. But every single one of them still made it. And surprisingly, their workflows weren't that special. They were just simpler and more clever than they seem. So, the first project that gained massive popularity despite being entirely vibe coded is Medvy. It's a healthcare platform with more than 500,000 active users. It covers a wide range of healthcare issues and provides not just tracking, but also 24/7 expert support. The story goes that Matthew Gallagher, who was working alone, used AI tools to build this app from end to end. The company pulled in $401 million in revenue in its first year and is on track to become a billion-dollar company within this year. Despite having no experience in coding, he was able to build this app using AI tools. He didn't rely on a single tool. He picked each one for its strengths. He used Claude and Grok models primarily for coding with ChatGPT as a secondary debugging tool. Midjourney handled image generation on the site and Eleven Labs powered the audio calls, which removed the need for human call support entirely. But coding tools alone don't run a healthcare company. So, instead of building pharmacies and shipping from scratch, he outsourced them to existing services. So, that took the burden of maintaining stock and delivering off from him. The same went for professional consultancy. He outsources the consultancy as well, eliminating the need to be involved personally in that aspect as well. He treated every dependency as a service, not a hire. His own job was product judgment, figuring out what the market actually needed. But running solo has a cost. One day he broke production while he was away. Nobody else could handle it and the company lost 200 customers in a single hour. Therefore, he hired two engineers, not to scale, as a safety net, so the next outage wouldn't repeat that same loss. The real skill here is being a better judge of what to build, which tools to assemble and when to stop. That comes from analyzing real user needs, not just collecting tools. Instead of building from scratch, he combined existing solutions in one place. And that's what actually brings customers and scales a company to a billion-dollar valuation. We share everything we find on building products with AI on this channel. So, if you want more videos on that, subscribe and keep an eye out for future videos. Now, Cal AI is a product that might sound like just another fitness tracker. But instead of manually adding the food you ate and the calories it contains the way normal trackers work, you can just upload an image of whatever you're eating and it converts that into calories and updates the database for you. It's available on both Android and iOS. It maintains a large database of foods and gives AI-powered suggestions, so you can monitor your weight and other nutrition goals in one place with ease. This product was built by two teenagers who were still in high school at the time, which then later scaled to more employees. It pulled in over 5 million downloads in just 8 months and generated over $2 million in revenue in a single month. It also held a strong 30% customer retention rate because most apps only gain users, but this one successfully retained them. It also holds a 4.8 rating on both the Play Store and the App Store. Now, this idea wasn't new. Similar apps already existed doing the same thing. But Cal AI had a real advantage the others didn't. It was built in the age of LLMs and used models from Anthropic and OpenAI to push accuracy up. It also relied on a large open-source food database and reached around 90% accuracy, which is more than enough for most diet enthusiasts. What really boosted this app wasn't heavy spending on marketing. It caught the attention of fitness influencers who played a major role in promoting it, leading to the spike in the users. Then we have Wave AI, which started with an idea so simple, yet it made a real impact on users. It's an AI-powered note-taking app that transcribes and takes notes for all kinds of meetings and recordings. Now, you might think there are already so many similar existing apps and the space for this is already so crowded. But Wave still broke through because it solves a problem people actually feel. During discussions, important details slip away and people need a reliable way to capture conversations across in-person and online meetings. It launched first as an iOS download, then scaled to Android and now it's available on every platform. The app was entirely vibe coded and pulled in around $7 million in revenue. The founder is not a developer at all, yet he scaled it into a million-making company. He ran the entire project completely solo. Similar to how Medvy operated, his infrastructure also leaned on third-party services instead of building everything from the ground up. He just integrated them into a friendly app and just focused on solving the problem in an interactive manner that made the user experience so much better. And this is what set this product apart from other similar existing ones. He used ChatGPT as his main tool and instead of asking it to build the whole app at once, he broke the application into smaller chunks. He prompted AI to write each part one by one. So, strategic positioning, focused user experience, and careful planning are what actually took him to that revenue level at speed. But before we move forwards, let's have a word by our sponsor, Scrimba. Most AI engineering courses are just someone talking over slides. However, with Scrimba, the experience is very different. Their web dev courses combine the video and code editor into one. Pause anytime, edit the instructor's code directly, and see what happens. No tab switching, no copy-pasting. That's how coding actually sticks. Their AI engineer path is what caught my attention. In under 12 hours, you go from zero to building real AI agents, learning rag, context engineering, and MCP, all built with JavaScript, so you don't have to learn Python first. They've even partnered with Mistral, LangChain, and Hugging Face to bring real tooling into the lessons. And beyond AI, they've got full career paths for front-end, full-stack, and back-end development with over 80 courses covering everything from React and Node to TypeScript and SQL. Use our link in the pinned comment to save an extra 20% on their pro plans. Get started today with their free courses and start building. Flypedia is another product built solely by AI, which started as just a fun hobby project, which then scaled to $500,000 a month. It is basically a browser-based flight simulator. He completely relied on AI tools for building and he was able to create the first version in just 30 minutes. The game scaled so fast that Elon Musk himself endorsed it. Its architecture was so well built that it was able to survive cyberattacks and started generating revenue at a serious scale. The whole thing was built using Cursor and it took the founder just 3 hours of Cursor to get the app about 80% done and in a state that it was ready to announce to the public letting them use it. His workflow itself was pretty simple. He started with one prompt and based on how the tool generated code and features, he iterated with new prompts. Each iteration added a feature or fixed an issue one by one layering in the game mechanics along the way. The game performed well if one person was playing, but scaling to multiplayer is where the project needed help. He was approached by the Beta List founder to help him fix the multiplayer issue by adding WebRTC, which solved to some extent, but it worked well only for two people. Therefore, the founder of Cursor himself reached out and they switched to WebSockets, which actually solved the problem and unlocked real-time multiplayer for everyone. He launched the game as a free version, but added a specific plane for $29. This helped it gain a lot of popularity and he made a significant amount of money in a short time. His stack was Cursor with Grok 3 as the back-end model, Claude Sonnet 3.7, and ChatGPT for debugging. He's just an indie hacker with no game dev background. What got him there was determination and a systematic step-by-step debugging approach. TrendFeed is another product that gained rapid popularity amongst the users and made solid revenue. It's a marketing tool aimed at content creators, focused on building and acquiring customers, growing a community around existing brands, and lifting overall revenue for the creators. The project pulled in around $12,000 in just 4 weeks. It was entirely built with AI using Cursor with Sonnet, not through Claude code, but directly inside Cursor. His build process was actually pretty straightforward. He started by analyzing the UI carefully and doing deep competitor research, even using AI to break down those competitors. Then he moved to data structure design, defining schemas with Cursor or Claude, and iterated from there. On launch day, the app generated 5.5 thousand pounds in a single day, which was a massive first-day result. Even though the founder is non-technical and works in fields outside computer science, he shipped the whole thing using AI. The app is built on Next.js, React, Shadcn, Superbase, and Vercel stack, all the tech AI tools work the best with. Given how popular the product became in such a short time, it was surprising that he spent zero on marketing. Instead, he leaned entirely on TikTok, Instagram, and YouTube to drive views and announce the product. His full build ran on Claude code and Cursor with Sonnet as the primary model. The flow itself was clean. He started with design, set up the core app structure, laid out onboarding and the main framework, and repeated design patterns. Then he broke the app into modular components that AI could build and merge together. Also, if you are enjoying our content, consider the hype button because it helps us create more content like this and reach out to more people. The next successful AI product that was entirely vibe coded is Aura. It is basically a site full of templates for beautiful websites with assets, components, and skills, all tailored towards strong design. The whole project was built by Meng To, who was the person behind Aura. He posted on X that the product hit $15,000 in monthly recurring revenue or MRR and gained over 21.7 thousand users in just a month. He also shared that he now uses Cursor for design and he's no longer using Figma like in his previous workflow. His main point is that you shouldn't just vibe code, you should also vibe design because AI tends to generate basic UIs. So, instead of letting it work on its own, you need to give it guiding templates to diversify the look. He recommends components from existing libraries like 21.dev. He also recommends not relying on a single model while building the app. Instead, it is more effective to start with Claude models because they are more powerful for coding tasks. And if it fails to do the task, then switch to Gemini or GPT models when needed. Instead of going all in at once, he stresses building the app step-by-step with incremental changes. He recommends keeping the prompt simple by breaking the app into smaller parts and iterating on them one at a time. He also says prompts should ideally stay under three sentences so the AI stays focused. You don't need to dump all the documentation on the AI either. Instead, you should give it the minimal but correct context so it delivers what you actually want. This way the agent will be able to focus more on the task at hand. In short, keep the agent setup simple and focused. Another product worth looking at is Sleek, which is a product that turns prompts into engaging websites. It generates the full design from a prompt, builds stunning visuals, creates mockups, and allows code export. The product reached $10,000 MRR in 6 weeks and was built entirely using AI tools. The impressive part is that the developers hit that MRR without spending a single dollar on marketing. But what really sets Sleek apart is that they didn't start from zero. They had already built other design tools before, so they essentially repurposed their existing products into this one. They used a stack of Next.js, Supabase, and Vercel, which AI tools already handle comfortably. They acquired all of their customers through X by making the use of its algorithm cleverly and announcing early access, which led to a strong launch. But here's the real reason the product succeeded. They had a clearly defined ideal customer profile or ICP from day one. Because of that, they understood exactly what their target users needed and could shape the product to fit. So whenever you build an app, define an ICP first. That's what separates successful apps from impressive ones that never make money. When your ICP is clear, you shape the product around a specific audience. Identify the right customer and build something they actually need and will pay for. And finally, there's Sideshore, another product built entirely with AI. It solved one of the biggest problems agents had at the time, which was hallucinating references, citations, and sources that turned out to be nonexistent when checked. It's a platform where you input citations and it verifies whether the AI-generated ones are actually correct. Even though it solved such a simple problem, it gained massive popularity. The site generated around $10,000 MRR and grew steadily. But the story doesn't stop there. The site was eventually acquired for a significant amount by Jenny AI, another AI-powered platform working in the same space. That makes it a strong example of how a simple but critical problem can turn into a valuable product. That brings us to the end of this video. If you'd like to support the channel and help us keep making videos like this, you can do so by using the Super Thanks button below. As always, thank you for watching and I'll see you in the next one.