Voice Assistant (Jarvis)

A Python-based voice assistant that uses speech recognition, text-to-speech, and an LLM to answer questions and execute commands.

Features

• Wake word detection ("Jarvis")
• Session-based listening
• Conversational memory for follow-up questions
• Command handling (open apps, play YouTube, etc.)
• Text-to-speech using Microsoft Edge TTS

Tech Stack

• Python
• SpeechRecognition
• Microsoft Edge TTS
• Gemini API

What I Learned

• Managing session state in voice systems
• Why LLM APIs are stateless and how to design memory
• Debugging timing and OS-level audio issues
• Structuring a project into clean modules

Limitations

• Microphone input is paused during speech to avoid audio conflicts on Windows
• Voice interrupt during speech is not supported on standard hardware

Status

This is a learning project and my first step into building AI-powered assistants.