Wattara

Wattara is an AI-powered household energy-planning agent designed for Peru, built using Muse Spark 1.3 through the Meta Model API. It translates natural-language electricity-saving goals into actionable plans by combining tool calling, simulated household consumption analysis, appliance-level anomaly detection, and LightGBM forecasting. Unlike a conventional chatbot, Wattara uses independent Python validation to recalculate savings, enforce protected-appliance restrictions, and require user approval before simulating any action. The prototype also compares predicted and simulated measured savings to evaluate the proposed plan. Wattara currently operates on synthetic household data without physical hardware. Future development will focus on real-world validation, personalized consumption patterns, and safe smart-device integration.