L.U.N.A.: Logical Universal Networked Assistant

AI-powered voice assistant with multimodal interaction

L.U.N.A. is an intelligent voice-controlled AI assistant that combines conversational AI, computer vision, and system automation into a unified desktop application. It enables hands-free interaction with your computer through natural voice commands, providing real-time information, automating tasks, generating images, and analyzing visual content from your screen or camera. The assistant features a modern PyQt5 interface with animated graphics and supports multi-turn conversations with context awareness. L.U.N.A. intelligently routes queries between offline processing and cloud-based AI services to balance speed, privacy, and capability.

Key metrics

  • Task Types: 10+
  • App Support: 100+
  • Response Time: <2s
  • Classification Accuracy: ~92%
  • Task Success Rate: ~95%
  • Voice Recognition: 85-90%

Tech stack

Python 3.10, PyQt5, Groq API (Llama 3-70B), Cohere API (Command-R-Plus), Google Gemini API (2.0 Flash), HuggingFace API (Stable Diffusion XL), Web Speech API, Edge TTS, asyncio, Selenium WebDriver, AppOpener, OpenCV, MSS (Screen Capture), Pillow, pygame, mtranslate, pywhatkit, googlesearch-python, BeautifulSoup4, dotenv

View on GitHub

← All projects