💡 Inspiration

I was "inspired" (read: deeply annoyed) by the bond calculators I used when my wife and I bought our house. Those tools serve exactly one purpose: to show a simple burndown at today’s interest rate. But that’s not how my bond was going to behave. In reality, I added money regularly, withdrew from the bond when I needed to, and the interest rate changed from time-to-time. How did that affect my loan term or monthly payments? The tools couldn’t tell me. I needed something visual, grounded, and responsive to real life.

🚀 What it does

Drift is a finance productivity app that helps you simulate and track how your debt changes over time. You can:

  • Model a loan and see how monthly payments evolve
  • Record changes like extra payments, skipped months, or rate adjustments
  • Run “what if” simulations to test different strategies — what if the interest rate jumps by 0.25%, or worse, 2.5%?

The goal is to make debt more visible, less abstract, and easier to manage.

🔧 How I built it

I leaned heavily into the AI-led approach—letting Bolt write most of the code while I focused on shaping the goals and behavior of the app. That’s the whole point of Bolt, after all.

The built-in documentation helped integrate tools like Supabase quickly. I also experimented with being intentionally vague in prompts to see how the AI interpreted high-level goals—and honestly, it worked. One early result was so spot-on, it changed how I thought about presenting Drift entirely. From that point on, the process felt more like creative collaboration than pure coding.

🧱 Challenges I ran into

Most of the challenges weren’t with the app idea itself—but with giving up control.

  • As an engineer, it was hard not to jump in and restructure the code. Letting go of that instinct was necessary to move fast.
  • Getting the AI to design UI the way I wanted took patience—sometimes it would add fields I didn’t ask for or ignore layout intent. I had to learn to prompt more precisely, and the Enhance Prompt feature helped a lot.
  • Finally, figuring out how to model debt changes in a way that’s both accurate and intuitive was trickier than expected. Real-world debt is messy, but the UI still needs to feel smooth and approachable.

🏆 Accomplishments I’m proud of

  • I shipped a working loan calculator prototype without getting bogged down in polish or tech debt.
  • I gave AI more control than I normally would—and it actually helped accelerate the work.
  • Simulations feel responsive and give useful feedback with minimal input.
  • I finally got this idea off the ground. When the hackathon launched, I knew it was the push I needed—and I didn’t waste it.

🧠 What I learned

  • Clarity of intent beats overly detailed specs—especially with AI.
  • Debt is messy, but visualization brings clarity. People don’t need perfect models; they need to see what’s going on.
  • You can ship real value without perfect code. Momentum matters more than maintainability in the early stages.

🔮 What’s next for Drift

  • AI-generated repayment strategies tailored to your actual loan behavior
  • Track changes to the prime interest rate and link them to flexible-rate debt accounts
  • Enable a “Partner Mode” so you can share accounts with a spouse or business partner
  • Long term: explore bank integrations for live debt syncing

Built With

  • react
  • react-charts
  • react-router
  • superbase
  • tailwindcss
  • vite
Share this project:

Updates