Inspiration

As a developer, I often find myself writing snippets of code—a clever function, a complex algorithm, or a new configuration—only to forget the context or the reasoning behind it weeks later. Code repositories save the "what," but they often lose the "why." I was inspired to create a tool that acts as an intelligent, personal code diary. A place where I could not only store my code but also get instant, AI-powered documentation and insights, creating a living "legacy" of my work that I could learn from in the future.

What it does

DevLegacy AI is a full-stack web application that serves as an intelligent code journal. A developer can log in securely, paste any code snippet, and receive an instant, in-depth analysis from the Google Gemini API. This analysis includes:

A plain English explanation of what the code does.

Actionable suggestions for improvement and best practices.

Automatically generated unit tests to ensure code quality.

Every analysis is automatically saved to the user's personal history, which is tied to their account. This creates a searchable, intelligent knowledge base of their own work, allowing them to revisit past solutions and learn from their own coding patterns.

How we built it

Building this application in just one weekend required a modern, full-stack approach.

The frontend is a responsive single-page application built with React and TypeScript, styled with Tailwind CSS and the Shadcn/ui component library for a polished, professional look.

The backend is a robust and secure REST API built with Java and the Spring Boot framework.

The core intelligence is powered by the Google Gemini API, which handles the complex task of code analysis and generation.

User data and all generated code analyses are stored in a MongoDB Atlas NoSQL database.

User authentication is handled securely by Auth0, providing a seamless and safe login experience.

Challenges we ran into

This hackathon was an intense and challenging journey, particularly in the realm of deployment and configuration. The biggest hurdle was a series of cascading dependency and environment issues. I battled through:

Dependency Hell: The initial frontend template had numerous conflicting dependencies, which required a complete, from-scratch rebuild of the frontend to resolve.

Deployment Nightmares: I faced platform-specific bugs on three different hosting services (Vercel, Render, and Fly.io) related to environment variables, build commands, and caching, which required deep debugging of each platform's unique configuration.

Authentication Flow: Implementing a secure, seamless login and redirect flow with Auth0, especially in a production environment, was a complex task that required multiple iterations to perfect.

Accomplishments that we're proud of

What we learned

This hackathon was an incredible learning experience. I didn't just build an app; I built resilience. I learned the critical importance of a clean project setup and the nuances of deploying a full-stack application. I gained hands-on, in-the-trenches experience with debugging complex issues across the entire stack, from frontend dependency conflicts to backend CORS and environment variable errors. Most importantly, I learned how to integrate multiple powerful APIs (Gemini, Auth0, MongoDB) into a single, cohesive product. This project pushed my limits and taught me that even the most frustrating bugs can be solved with patience and a clear strategy.

What's next for Devlegacy Ai

Share this project:

Updates

Submission history