Inspiration
The inspiration behind AppSparks came from seeing how difficult it can be for beginners to learn programming. Many students and innovators have great ideas but get stuck at the coding stage. I wanted to build something that lowers the barrier to entry and helps anyone turn their imagination into working code instantly.
What it does
AppSparks is an AI-powered tool that converts plain English ideas into runnable code. Users simply type what they want—like “build a calculator” or “make a to-do app”—and AppSparks generates the code instantly. It also saves past queries in a history, so users can revisit, reuse, or improve their earlier projects.
How we built it
We built AppSparks using: Python (Flask) for the backend Google Gemini API for AI-powered code generation SQLite with SQLAlchemy for storing query and code history HTML + Jinja2 templates for the frontend interface VS Code as the development environment
Challenges we ran into
Initially tried using OpenAI API but ran into billing and quota issues. Switching to Google Gemini required adapting the codebase quickly. Debugging HTML templates (like history.html) took time and effort. Time limitation (2 days) made it important to focus only on the core features.
Accomplishments that we're proud of
Successfully built a full working prototype in a very short time. Integrated AI with a database and frontend to create a complete product. Created a tool that is useful for both beginners learning to code and developers who want to prototype faster.
What we learned
How to integrate and adapt different AI APIs (OpenAI → Gemini). Building a full-stack application that connects frontend, backend, and AI. Importance of error handling and fallbacks when working with APIs. That sometimes focusing on simplicity leads to the best outcomes.
What's next for AppSparks
Expand beyond Python to support more languages like JavaScript, C++, or Java. Build a web-based platform where people can share their generated projects. Add features like debugging assistance, code explanations, and one-click deployment. Scale the project for real-world use, not just a prototype.
Log in or sign up for Devpost to join the conversation.