N.A.T. AI – Intelligent Personal Assistant
Category
Full-Stack
Tech Stack
Python, FastAPI, Groq AI, LangChain, LLaMA, FAISS, Playwright, TTS, Voice AI, WebGL, Vanilla JS, AI Assistant
Featured
No
Live URL
N/A
Overview
N.A.T. (Natasha) is a powerful local AI assistant built with Python FastAPI and Groq AI (LLaMA 3.3 70B). It features three chat modes — General, Realtime (web search via Tavily), and Intelligence (market analysis). The system includes a Smart Action Engine that classifies user intent and routes to the correct service: terminal commands, browser automation via Playwright, app launching, or conversational AI. It supports streaming responses via Server-Sent Events (SSE), real-time text-to-speech using Microsoft Edge TTS, and browser-native voice input. A FAISS vector store with HuggingFace sentence-transformers powers a personal learning system from local text files. The frontend is a vanilla HTML/CSS/JS dark glass-morphism UI with an animated WebGL orb that reacts to speech. No build tools required — everything runs on a single Python server.