🧠 Voice-Assistant

• Voice-Assistant is a ML-based Voice Assistant

• This is a desktop-based voice assistant powered by the ML model, which is built using Python.

• The UI is built using Electron.js while inference is done by using Node.js.

• It understands your speech, processes it using a custom-trained intent recognition model and executes platform-specific commands.

---

🚀 Features

• 🎙️ Voice Input

• 🤖 ML-based Intent Detection

• ⚙️ Cross-platform Command Execution

• 🗣️ Voice Output

• 🖥️ Desktop-first App

• 🔐 Privacy-Focused

---

🏗️ Tech Stack

Layer	Stack
Frontend	HTML, CSS, JS and React.js
Backend	Node.js and Python
Model	Sci-kit Learn ML models
Voice I/O	Web Speech API

---

🧠 How It Works

1. 🎙️ User speaks a command

2. 🎧 Voice is transcribed and sent to Python backend

3. 🧠 The ML model classifies intent

4. 💻 Executes the platform-specific command

5. 🗣️ Response is spoken back to the user alongwith the execution of the task

---

🔁 Requirements

• Node.js & npm

• Python & pip

• Electron.js & React

---

🛠️ Run Locally

1. Clone the repository

git clone "https://github.com/hrutavmodha/voice-assistant.git"
cd ./voice-assistant

2. Train the models

cd ./backend/model/model1
python ./train.py

cd ../model2 && python ./train.py

cd ../model3 && python ./train.py

cd ../model4 && python ./train.py

3. Install dependencies

cd ..
npm install
pip install sklearn joblib 
cd ../frontend
npm install
cd ..
npm install
cd backend

4. Start the main server

npm run start

5. Start the voice assistant's server

python -m frontend.chat.app

---

👨‍💻 Author

Made with passion by Hrutav Modha

---

📌 Note

• After running the app.py, wait for some time, the browser will also get opened automatically at the URL http://localhost:3000

• This project is purely offline and does not rely on any of the external APIs and nor it is dumb if-elif-else-like rule-based voice-assistant

• It’s a fully local voice assistant with a custom-trained AI model, which is designed with keeping privacy in mind.