Desktop Voice Assistant

A Python-based desktop voice assistant for hands-free system interaction, application automation, and voice-driven desktop workflows.

---

Overview

The problem: Desktop automation typically requires manual input — clicking, typing, navigating menus. Voice-driven interaction can streamline common tasks, but most voice assistants are cloud-dependent or tied to specific ecosystems.

The approach: A lightweight, offline-capable desktop assistant that combines speech recognition, natural language command parsing, and system automation. The assistant listens for voice commands, interprets intent through keyword matching, and executes actions — from launching applications to controlling browser workflows.

Architecture: Four-module pipeline — speech recognition (speech_to_text.py), command routing (action.py), speech synthesis (text_to_speech.py), and the main orchestration layer (Allen.py) with a Tkinter GUI.

---

Features

• Voice command processing — Real-time microphone input with speech-to-text conversion
• Application launching — Open Chrome, YouTube, Google Maps, and other tools by voice
• Browser automation — Navigate to Spotify, WhatsApp, Gmail, Google Calendar, Zoom, and more
• System information — Voice-query current time
• Text-based input — Alternative keyboard input mode alongside voice
• Conversation history — Scrollable chat interface showing user and assistant messages

---

Technology Stack

Technology	Purpose
Python	Core language
SpeechRecognition	Speech-to-text via Google Web Speech API
pyttsx3	Offline text-to-speech engine
Tkinter	Desktop GUI framework
Webbrowser	System browser integration

---

Getting Started

Prerequisites

• Python 3.7 or later
• Microphone (for voice input)

Installation

# Clone the repository
git clone https://github.com/Ajaykumar-Mallameeda/Desktop-Voice-Assistant.git
cd Desktop-Voice-Assistant

# Install dependencies
pip install -r requirements.txt

# Run the assistant
python Allen.py

Dependencies

See requirements.txt for the full list:

• speechrecognition — microphone input and speech-to-text
• pyttsx3 — text-to-speech synthesis
• pyaudio — audio I/O for microphone access
• Pillow — image handling for GUI
• pipwin — Windows audio dependency helper (Windows only)

---

Project Structure

Desktop-Voice-Assistant/
├── Allen.py                 # Main entry point and GUI
├── action.py                # Command routing and execution
├── speech_to_text.py        # Speech recognition module
├── text_to_speech.py        # Speech synthesis module
├── .gitignore
├── LICENSE
├── README.md
└── requirements.txt

---

Usage

Run the assistant:

python Desktop\ voice\ assistant/Allen.py

The GUI window opens with an ASK button (voice input), a SEND button (text input), and a DELETE button (clear chat).

Example commands:

• "What is your name"
• "What is the time"
• "Open YouTube"
• "Play music allen"
• "Open Google"
• "Open WhatsApp"
• "Shutdown"

---

Known Limitations

• Hardcoded image path in the GUI — currently points to a local file that doesn't exist on other machines
• Module-level function calls in speech_to_text.py and text_to_speech.py execute on import
• Single-threaded — voice recognition blocks the UI during processing
• Keyword-based command matching (no NLP/LLM integration)

These are documented as areas for future improvement.

---

Lessons Learned

• Separation of concerns — Splitting STT, TTS, and action routing into separate modules made the system testable and extensible. Adding new commands requires only editing action.py.
• Offline vs. cloud trade-offs — Using pyttsx3 for offline TTS (no internet required) meant lower voice quality but higher reliability. The speech_recognition library uses Google's cloud API, introducing a network dependency for input.
• GUI frameworks matter — Tkinter is built-in and lightweight but limited in modern UI capabilities. A web-based interface would enable richer interactions.

---

License & Author

License: This project is licensed under the MIT License — see the LICENSE file for details.

Author: Ajaykumar Mallameeda · Indian Institute of Technology Palakkad

---

Built at IIT Palakkad as part of a continuous learning journey in AI and Backend Engineering.