CurateBang

logo

About the project

What it does

CurateBang is a minimalist, AI-powered content amplification engine. It solves a major pain point for creators by transforming a single idea or prompt into multiple pieces of high-quality, tailored content for every major platform—from a professional LinkedIn article and an insightful blog post to a punchy Twitter thread. Its clean, distraction-free interface is designed for one purpose: to help creators work smarter, not harder.

Inspiration

As a developer and creator, I was constantly frustrated by the repetitive, time-consuming process of manually adapting my ideas for different social media platforms. I knew there had to be a better way. I was inspired to build a tool that would not only automate this process but do so with an elegant, minimalist user experience. The goal was to create a platform that felt like a true partner in the creative process, amplifying ideas without adding complexity.

How I built it

CurateBang is built on a modern, fully serverless architecture, with every component chosen for performance, scalability, and developer experience.

AI-Powered Development with Kiro: The entire project was architected and built using Kiro. I relied heavily on spec-driven development to define the core features, which allowed Kiro to generate robust, maintainable code for both the frontend and backend. This AI-assisted workflow was instrumental in accelerating the development process and allowing me to focus on high-level architecture and user experience.
Frontend: The user interface is a fast, responsive Next.js application, built with TypeScript and styled with Tailwind CSS.
Database: For the database, I chose NeonDB, a serverless Postgres platform. This provided the power and familiarity of SQL with the scalability and ease of use of a modern, serverless architecture.
Hosting: The application is hosted on Vercel, enabling continuous deployment and a globally fast experience for all users.
AI Engine: The core content generation is powered by the Perplexity API, with serverless functions handling the secure API calls and prompt engineering.

Challenges I ran into

The primary challenge was architectural. I initially started building the backend with Convex, but as the project's data models grew, I realized that a relational database would be a better fit. Pivoting to NeonDB was a significant decision that required refactoring the backend logic. While challenging, this move ultimately resulted in a more robust and scalable architecture.

Another interesting challenge was prompt engineering. Fine-tuning the prompts sent to the Perplexity API to generate consistently high-quality, platform-specific content required significant iteration and testing.

What I learned

This hackathon was an incredible learning experience. I deepened my understanding of modern, decoupled web architectures and the power of a fully serverless stack. Most importantly, this project was a masterclass in AI-assisted development. Using Kiro for spec-driven development wasn't just a novelty; it fundamentally changed my workflow, proving that AI can be a true partner in building high-quality, complex applications faster than ever before.

What's next for CurateBang

The vision for CurateBang is to become the go-to content co-pilot for creators. Future plans include adding more platforms, integrating with social media scheduling tools, and introducing more advanced AI features, like the ability to learn and adapt to a user's unique writing style over time.