A 3rd year CS student's attempt to reduce AI's water footprint — EcoCache (A Python Library)

A 3rd-year CS student created EcoCache, a Python library that caches LLM API responses using semantic similarity (sentence-transformers + FAISS) to avoid redundant queries. It reports ~5 mL water and ~4 g CO₂ saved per cache hit, with a live dashboard tracking cumulative savings. The project is open-source (v0.1) and aims to reduce the environmental cost of repetitive AI usage at scale.