LLM Lite: The AI Cost Optimizer

Inspiration As a high school student experimenting with AI, I realized API costs were a massive barrier for startups. I wanted to build a "Sustainability Layer" for AI that saves both money and compute resources. What it does LLM Lite acts as a smart proxy between the user and LLM providers. It caches frequent queries (reducing redundant calls) and routes simpler tasks to smaller, cheaper models, optimizing cost without losing quality. How we built it I used Python/FastAPI for the backend to ensure low latency. The routing logic was "vibe coded" using AI assistance to iterate quickly on the best algorithms for token estimation. What's next for LLM Lite I plan to add a "Carbon Footprint Tracker" to show users how much energy they saved, aligning with Green AI principles.