JARVIS Voice Assistant

Project Overview

An AI-powered voice assistant capable of natural voice interaction, query resolution, and local/cloud automation. It operates directly at the OS level to automate tasks, run files, check emails, control browsers, and execute system commands.

Development Process

Developed in Python using Google Speech Recognition APIs and Pyttsx3 text-to-speech engine. Built a modular command parsing architecture utilizing regular expressions and basic NLP. Connected local llama-based LLM APIs to handle contextual memory, maintaining conversational state between turns.

Key Features

• Real-time voice query recognition and text-to-speech synthesized replies
• System automation including app launching, file system editing, and media controls
• Web automation using Selenium and Web scraping for automated search and retrieval
• Intelligent memory module to store preferences and contextual user history
• Custom trigger word detection and background microphone loop execution

Challenges & Solutions

Synthesizing text-to-speech asynchronously without blocking the microphone loop. Resolved by implementing Python's threading library to decouple command execution and speech processing, keeping the microphone listener active.

Technologies Used