An AI prompt optimizer that rewrites verbose prompts for efficiency and visualizes their environmental impact — built in 24 hours at FullyHacks 2026.
AI prompts are often unnecessarily verbose and computationally wasteful, with users having no visibility into the energy cost of their AI usage.
A web app that rewrites prompts using a local Ollama LLM pipeline, then scores and visualizes the environmental impact reduction via the HumanDelta metric.
Built the full stack in 24 hours — Next.js frontend with Tailwind, FastAPI backend, Ollama LLM integration, and PostgreSQL via Neon for persistence.