This is a submission for the World’s Largest Hackathon Writing Challenge: Building with Bolt. TL;DR: I built a mobile app called RideLink that compares ride-hailing prices across multiple services (Uber, Lyft, and more) and launches the chosen ride with pre-filled details, saving time, money, and taps.
I did this in one month using Bolt.new’s AI-powered development platform, which transformed how I approached coding and helped me go from a personal pain point to a working prototype fast.
The Idea Born from a Personal Pain Point
The inspiration for RideLink came from my own frustration. I often found myself juggling between Uber, Lyft, and other ride apps to check which had the cheapest fare or the quickest arrival. Ride prices vary wildly across apps, yet there was no quick way to compare them in one place. I’d manually copy and paste addresses from one app to another, a tedious process that begged for a solution. When the World’s Largest Hackathon was announced, I knew it was the perfect opportunity to turn this pain point into a project. RideLink’s concept was simple: allow users to input a destination (and optionally a pickup location), then display fare or price estimates from multiple ride services side by side. With one tap, the user could launch their chosen ride app with the pickup/dropoff already filled in. In short, it’s a smart ride-launcher that puts the power of choice in the rider’s hands without the usual app-switching hassle. I envisioned that RideLink could save people time, money, and a lot of screen-taps every day.
Adopting an AI Co-Developer with Bolt.new
I had never used a tool quite like it before, and I was eager to see how it could accelerate the creation of RideLink. Using Bolt felt like pair-programming with an incredibly fast, tireless junior developer. I would describe a feature or ask a question in natural language, and Bolt’s AI would generate code, suggest libraries, or outline how to implement the feature. For example, early on I prompted Bolt with something like: “Create a simple React Native screen with two input fields (pickup & drop-off) and a button that compares prices for Uber and Lyft.” In seconds, Bolt provided a starter code snippet using React Native and even suggested integrating a mapping API for address autocomplete. This initial boost was invaluable—it gave me a structured starting point that I could then refine and expand. Bolt.new transformed my development process by handling a lot of boilerplate and repetitive tasks, which freed me to focus on the core logic and user experience. Instead of wrestling with setting up navigation or state management from scratch, I could ask Bolt to scaffold it. When I needed a new component or encountered a bug, Bolt was there 24/7 to offer potential solutions. It was as if I had an on-demand coding assistant who never slept!
Technical Journey and Breakthroughs
Building RideLink wasn’t a straight line, it was a journey filled with technical challenges, aha! moments, and learning experiences. Here are some key aspects of the development process and how I tackled them (often with Bolt’s help):
Multi-Service Deep Linking: The core feature of RideLink is launching different ride apps with pre-filled route information. Each ride-hailing app (Uber, Lyft, Bolt (the ride service), etc.) has its own URL scheme and parameters for deep linking. This was initially daunting, I had to ensure that when a user selects a service, the app opens correctly with the pickup and drop-off already set. For instance, Uber might require coordinates and an action parameter, while Lyft might use a different URL format entirely. I leaned on Bolt’s AI to help generate and verify these URI schemes. I’d ask Bolt things like “How do I open the Uber app to request a ride from code?” and it returned the format (uber://?action=setPickup&pickup[latitude]=… etc.) which I then adjusted for our app. Getting each app’s deep link format right was a challenge, but with iterative prompts and some manual tweaking, I managed to implement reliable links for all supported services. The moment I first tapped “Request Ride” in RideLink and saw Uber and Lyft pop open with the addresses already filled was pure magic, seeing it work for the first time was a moment of pure joy!
Smooth Address Autocomplete: To make RideLink user-friendly, I needed an address autocomplete for entering destinations (and pickup, if not using current location). I integrated the Google Places API via an Expo library to provide place predictions as the user types. The technical hurdle here was ensuring the autocomplete was fast and didn’t bog down the app. Bolt’s suggestions pointed me to use debouncing (to limit API calls as the user types) and how to properly secure the API key. Implementing this, I learned how to manage API calls and state in a React Native app, and Bolt assisted by generating an example using the GooglePlacesAutocomplete component. After some fine-tuning, RideLink could suggest addresses nearly as smoothly as the native ride apps do, a big win for the user experience.
State Management for Preferences: RideLink isn’t just about one-time comparisons; I also added features like favoriting certain locations (home, work, etc.) and toggling which ride services to include in comparisons. This meant I had to manage state for these preferences and persist them. As a newcomer to React Hooks and state management, I found this part educational. I used the Context API for global state (to handle which services are toggled on/off across the app) and AsyncStorage for persistence, so your preferences would be saved between sessions. Setting that up was made easier by Bolt.new, for example, I asked it “How to store user preferences locally in React Native?” and it guided me toward AsyncStorage with sample code. There was a tricky moment when I noticed toggling a service didn’t immediately reflect in the UI. I realized I needed to use state update functions properly to ensure the UI re-rendered. It was a crash course in React reactivity, but once solved, the app felt much more responsive and robust.
Cross-Platform UI & Expo: I chose React Native with Expo to build the app, meaning it had to run nicely on both Android and iOS devices. Expo’s tools are fantastic, but I did encounter platform-specific quirks (like differences in how deep links are handled or subtle layout differences between OSes). Using Bolt, I was able to quickly get solutions for these quirks. One example: on one platform, the map view that shows the route preview was not resizing correctly. I described the issue to Bolt (“Map view not resizing on Android when keyboard opens”) and it pointed me to a known issue and a workaround involving KeyboardAvoidingView. Moments like that saved me from hours of googling through random forum posts. The result was a clean, intuitive interface that truly felt native on both platforms, something I’m proud of as a designer-developer.
Location Permissions & Default Pickup: For convenience, RideLink can use your current location as the default pickup point. Handling location permissions can be tricky, if the user denies access, the app should handle it gracefully. I used Expo’s Location API for this. Bolt helped by providing a quick snippet on requesting permissions and fetching the current GPS coordinates. I implemented a fallback so that if permission is denied, RideLink simply asks the user to manually enter a pickup address (instead of crashing or hanging). This was one of those polish points that I might have overlooked if I weren’t thinking holistically, but hackathons encourage you to consider edge cases too. It felt great to get this right: no matter if the user grants location or not, the app still works without a hitch.
Throughout these technical hurdles, AI assistance was a theme. Bolt.new, along with occasional help from ChatGPT for general questions, dramatically accelerated my problem-solving. That said, it wasn’t a magic wand, I often had to debug, refine suggestions, and combine features in ways the AI didn’t fully predict. For example, merging the Google Places autocomplete with my state management required manual adjustment. However, the AI got me ~80% of the way there each time, and that was huge given our tight one-month timeline.
How AI-Powered Development Changed My Approach
Working with Bolt.new not only sped up development, it fundamentally changed how I approach building software. In the past, I might have been hesitant to dive into a full-stack mobile app project all by myself, anticipating weeks of setting up infrastructure and countless trips to documentation pages. With an AI co-developer by my side, I felt empowered to attempt things outside my comfort zone because I had a safety net. When I got stuck, I’d just describe the problem or the bug in plain language, and more often than not, Bolt would unearth a clue or solution I hadn’t considered. I found myself focusing more on what I wanted to build rather than worrying about how to code every little piece from scratch. This shift in approach was liberating. For instance, I could concentrate on the user flow of comparing ride prices, the logic, the design, the value to the user, while Bolt handled the syntax of implementing that logic in React Native. It was like I could transfer some mental load to the AI and allocate more of my brainpower to higher-level problem solving and creative work. There were also moments of learning where Bolt’s output became a tutor for me. If it generated code using a library or function I wasn’t familiar with, I’d take a moment to understand that piece. In this way, I was learning by doing with AI guidance. Over the month, I became noticeably more comfortable with technologies that were new to me at the start (such as Expo Router for navigation and various native modules). The next time I build an app, I know I can reuse a lot of that knowledge and I’ll certainly feel more confident leveraging AI tools to boost my productivity.
Final Thoughts
In just one month, RideLink went from a back-of-the-napkin idea to a functioning prototype, thanks in large part to Bolt.new’s AI-driven development support. The technical journey was intense and incredibly rewarding. I not only ended up with an app that scratches my own itch (and hopefully will help others), but I also leveled up as a developer. The process showcased the power of combining human creativity with AI assistance: I brought the vision and user perspective, and Bolt brought the speed and coding firepower. For anyone hesitant about using AI in development, my experience is a testament to how game-changing it can be, especially when you’re a small team or solo builder trying to accomplish a lot in a short time. Building with Bolt has forever changed my dev toolkit and approach. I’m walking away from this hackathon with a project I’m proud of, a host of new technical skills, and a reinforced belief that no idea is too ambitious when you have the right tools. RideLink started as my personal remedy for ride-app fatigue, now it’s a real app that lives on my phone, built in a fraction of the time it would have taken me otherwise. Special thanks to my teammates Chris and Zack. That feeling of creating something tangible and useful is why I love building things. And with AI by my side, I’m excited to keep pushing the boundaries of what I can build next.