💡 Inspiration

LuminaSpace AI was born out of a common frustration: visualizing interior design changes, furniture placement, and room transformations is often expensive, slow, and heavily reliant on professional software that lacks real-time creativity. We wanted to bridge the gap between imagination and reality, empowering anyone to instantly redesign their living spaces using cutting-edge artificial intelligence.

🚀 What it does

LuminaSpace AI is a next-generation web application that turns any standard room photo into a beautifully staged, custom-furnished living environment. Users can upload room snapshots, select specific interior design styles, and leverage advanced AI generation to visualize custom renovations, optimize layouts, and seamlessly interact with a shop-the-look smart cart experience.

🛠️ How we built it

Frontend & UI: Built using React, Vite, TypeScript, and Tailwind CSS for a high-performance, responsive user interface.
AI Integration: Implemented direct client-side model architectures utilizing the Hugging Face Inference API with robust multi-key rotation and fallback chains (supporting models like SDXL, SD3, and FLUX).
Architecture & Reliability: Refactored to a resilient client-side execution model ensuring lightning-fast rendering, seamless image processing, and full zero-downtime reliability.

🧗 Challenges we ran into

During the final deployment phase, we faced a major architectural roadblock: serverless function limits causing 404 connection errors on backend endpoints during heavy image-generation calls. Overcoming this required a complete refactoring of our system architecture under strict time constraints—migrating all AI calls to a secure, robust client-side routing and fallback service.

🏆 What we've learned

This project pushed our technical boundaries, teaching us deep lessons in resilient software architecture, asynchronous client-side AI handling, API key rotation strategies, and maintaining absolute stability under pressure during a hackathon lifecycle.