Hugo is a Node.js-powered, modular voice assistant, with an easy way to add new features!

Note: looking for testers for macOS!

Features

• Can run completely offline
• Very customizable
• No 3rd-parties can access your data
• Compatible with Linux (tested), Windows 8+, and macOS

Installation

1. Clone the repository:

   git clone https://github.com/CharlesBobOmb/HugoVoiceAssistant.git
   cd HugoVoiceAssistant
   

2. Download an appropriate speech recognition model from here (recommended model is vosk-model-small-en-us-0.15)

3. Extract the contents into a folder called vosk-model inside of core/speech-recognition

4. [Windows/Linux] Install espeak-ng (for text-to-speech) from this repository

5. Install NPM dependencies and generate intents:

   npm install
   npm run generate-intents
   

6. Start Hugo by running:

   node index.js
   

...and you're done! Congrats!

To test Hugo, make sure a microphone and speaker are connected, run the program, and say "Hugo". Once a sound is played, ask Hugo "what time is it?" If everything is set up right, Hugo will tell you the time.

Changing Hugo's name

You can also change Hugo's name by adding a WAKE_WORD property to a .env file in the root. Make sure that the wake word is contained in your Vosk model (by using the say module), otherwise Hugo won't wake up!

Modules

What are modules?

Modules are code files that can be added to Hugo to let him do more things. Every module contains