VoiceAssistant is a resource-efficient and customizable voice assistant that is still in the early stages of development but already functional.

Privacy is our top priority, and VoiceAssistant ensures that all operations are performed offline on your local device, keeping your data secure and protected.

Goal

The main goal of VoiceAssistant is to provide a fully customizable and extendable voice assistant that can be controlled entirely using voice commands.

Features

Fully customizable and extendable (I'm still working on this)
Offline functionality for enhanced privacy and security
Voice recognition for executing predefined commands
Privacy-first approach with everything offline and local on the device

Supported Platforms

The project has been tested on x86_64 Linux, but it should theoretically work on the following platforms:

[x] Aarch64/Arm64 Linux/Android/Raspberry Pi
[x] Armv7l Linux/Android
[x] Riscv64 Linux
[x] x86 Linux/Android
[x] x86_64 Linux/Android
[x] 32-bit Windows (not suggested and untested)
[x] 64-bit Windows
[x] macOS (Intel)
[ ] macOS (M1)
[ ] iOS
[ ] WebAssembly

Note: The Android build works and starts, but I can't load the speech-to-text plugin.

To-Do's

[ ] Full customizability
[ ] Implement Whisper.cpp backend (this will add WebAssembly, iOS, and M1 macOS support)

Customization

In the near future, I'll add settings to the app!

To customize the commands, edit the commands/<your_language>/default.json file.

There are some default commands. See Add Commands.

Add Commands

To add your command, edit the commands/<your_language>/default.json file. The supported values are:

Name	Type	Description	Notes
`commands`	Array of strings	All commands that the voice assistant should react to	Required
`funcName`	String	The name of the function to call (`MainWindow::<funcName>`)
`responses`	Array of strings	It selects a random response from the array	Use `[wait <number of ms>]` to wait between words
`program`	String	A program to execute
`args`	Array of strings	Arguments passed to `program`	`${TEXT}` will be replaced with the voice input
`sound`	String	Path to local or remote file that gets played	Volume can be controlled via voice/GUI

Build

See the build section in INSTALL.md (direct link).

Translation Status

You can check the translation status of VoiceAssistant on Weblate (click the image below):

Creating a plugin

See plugins/README.md

Credits

VoiceAssistant uses the following open-source projects:

Vosk for voice recognition
11Zip for unzipping voice models
JokeAPI for telling jokes

Thank you to these amazing projects and their contributors for making VoiceAssistant possible!