Jarvis AI Assistant

This is a voice-activated AI assistant powered completely offline:

• Ears (Offline STT): Uses OpenAI's Whisper (tiny.en model) to listen to you without sending data to the cloud.
• Brain (Offline LLM): Uses Qwen2.5-0.5B-Instruct (a small, fast, intelligent local model) to think and converse.
• Voice (Offline TTS): Uses Audio8/Audio8-TTS-Preview-0.6b to speak back to you with a realistic voice.
• Antigravity CLI Support: Can execute agy commands and run system tools directly on your PC when requested!

Setup

1. Activate the Virtual Environment:

   cd C:\Users\AHMED\jarvis_assistant
   .\venv\Scripts\activate
   

2. Run Jarvis (Fully Offline):

   python jarvis_offline.py
   

Note on First Run

The first time you run jarvis_offline.py, it will download the Whisper model (~150MB), Qwen LLM (~1GB), and Audio8 TTS model (~1.2GB) from Hugging Face. Subsequent runs will be completely offline and much faster.

Usage

Simply speak into your microphone when the console says "Listening...".

• Try saying: "Jarvis, run the antigravity command agy status."
• You can say "quit", "exit", or "shut down" to close the assistant.