Project Information

  • Name: Sarthi (Virtual Assistant)
  • Date: 31 Jan 2024
  • Goal: Desktop Automation via Voice
  • Tech Stack:
    Python (OOP) Speech Recognition gTTS Threading
  • Modules:
    webbrowser playsound pywhatkit

Project Demo

Overview

We have all dreamt of creating our own AI assistant, inspired by movies like Iron Man (JARVIS). I developed Sarthi, a smart program designed to automate simple desktop tasks using voice commands. It serves as a hands-free bridge between the user and the operating system.

Capabilities

Play Songs: Stream music directly from the device or online.
Send Messages: Automate WhatsApp messages to saved contacts.
Web Search: Perform hands-free searches on Google & YouTube.
Camera Control: Open webcam and capture photos.
Note Taking: Dictate and save notes to local files.

How it Works

The core logic utilizes a JSON-based token system. Instead of hardcoding every possible sentence, the system maps specific keywords (tokens) to function pointers in a configuration file.

This architecture makes the assistant:

  • Simple: Easy to add new commands by editing the JSON file.
  • Fast: O(1) complexity for command lookup using Hash Maps.
  • Scalable: Logic is separated from the command definitions.