Library AI Platform

This project is a full-stack AI product featuring a clean, modern interface for engaging with AI models uniquely fine-tuned on selected knowledge sources, from books to podcast episodes. Users can explore a dynamic library, start contextual conversations, manage chat histories, and personalize their experience through account settings, all within a lightweight, scalable, and future-proof architecture.

Creating a system that handles unique AI models per source, supports rapid scaling to various platforms, and offers an elegant user experience while maintaining performance, modularity, and developer clarity was a significant design and engineering challenge. Cost was a key challenge as the budget to build the product was zero. Finding a cost-effective solution while maintaining quality was a major challenge.

I built the entire system with no budget, so every decision had to be efficient. I used Supabase for the backend because its free tier covers everything I needed: authentication, database, and storage. For the AI models, I used Gemini 2.0 since it was free and powerful enough for contextual chat. To move fast, I relied on AI tools to help write boilerplate, generate UI components, and streamline development. The architecture is clean and future-proof, designed to work seamlessly across platforms, with Expo-ready structure for fast mobile app rollout. I ensured that everything from the chat framework to the library filtering is built to balance usability with performance and scale.

The system architecture and component design enable rapid deployment with no restructuring required. The clean UX has been praised for its performance and design, while the underlying model switching architecture supports future expansion to thousands of source-based LLMs. The product is positioned as a next-gen AI knowledge companion platform.