Voxa is a voice-first AI-powered productivity assistant that transforms spoken input into structured daily plans. Designed for users who prefer talking over typing, Voxa uses speech-to-text processing, LLaMA 3.1 for intelligent task generation, and a Retrieval-Augmented Generation (RAG) system to provide personalized, context-aware planning.
Users can speak freely about their day, and Voxa will interpret the input, break it into prioritized tasks, estimate durations, and even read the plan back to the user using text-to-speech. With optional memory of past tasks and plans, Voxa acts like a daily productivity companion that adapts over time.
The project includes a React.js frontend and a FastAPI backend, integrated with Ollama for local LLaMA inference and ChromaDB for vector-based memory and retrieval.