Ameli-AI

*NOT READY FOR PRODUCTION/DAILY USE*

Welcome to Ameli-AI, a simple, beginner friendly but very ambitious voice assistant

Our goal is to run Ameli-AI in multiple devices such Linux, Windows, macOs, Android and iOs

Built with love and python3.10 Based on Knick-AI

Features

• Cross platform (currently linux and partial windows support)
• Features Google for online speech recognition
• Vosk and pytssx3 for offline speech recognition
• Online text-to-speech with GoogleTTS
• Offline text-to-speech with RHVoice (Brazilian-Portuguese, Russian and Ukrainian voices)
• Cross platform (Linux/Widows/macosx, support for android and ios also planned)
• Customizable Wake-word (like "OK Google" or "Hey Siri") support
• Integration with Google APIs like Google Calendar and Google Contacts
• Replaced keyboard module (used by Knick-AI) with pynput
• Replaced OpenWeatherMap (used by Knick-AI) with OpenMeteo, thus avoiding the use of an API key

Commands implemented so far

Feature	Description	Linux	Windows	macosx
WakeWord	Enables you to use your own customized hotword like "OK Google" or "Hey Siri"	X	X	X
AppFinder	Lets you find your installed apps and perform actions with them.	X	Partial	Partial
GoogleTTS (online text-to-speech)	Use Google's TTS Engine. Check supported languages.	X	X	X
RHVoice (offline text-to-speech)	Use RHVoice TTS Engine. Check supported languages.	In progress
Vosk (offline speech recognition)	Use Vosk for speech recognitions. Check supported languages.	In progress
(Partial) mpris support	Currently, allows you to stop any compatible media player with mpris protocol.	X
Browser (like "open netflix")	Enables you to open multiple websites. NavMode under progress (will take some time, might not even work :( ))	X	X	X
Take screenshot	Uses python to take screenshots of your screen.	X	X	X
Humanized time/date questions (like "what weekday is today", "what year is this")	Uses arrow to give you more humanized answers. Check available options.	X	X	X
Computer actions	Supported actions: reboot, shutdown, logoff and open terminal / console /command prompt	X	X	X

Available add-ons so far

Feature	Description	Linux	Windows	macosx
LofiGirl*	Allows you to focus better on your tasks when you're inside a noisy environment	X	X	X

*I had to disable the original Lofi Girl extension I've done because it was getting video directly without ads, not allowing creators to monetize their content.

Installation

Install on linux with git:

sudo apt-get install python-dbus libgirepository1.0-dev gcc libcairo2-dev pkg-config python3-dev gir1.2-gtk-3.0 libplayerctl-dev libdbus-1-dev libdbus-glib-1-dev python3-pyaudio libcairo2-dev pkg-config python3-dev
git clone https://github.com/asdf-vm/asdf.git ~/.asdf --branch v0.10.0
asdf=/home/$USER/.asdf/bin/asdf
asdf update
asdf install python 3.10.5
asdf global python 3.10.5
asdf reshim python 3.10.5
pip install pipenv
asdf reshim python 3.10.5
asdf global python 3.10.5
git clone https://github.com/andriusluz/ameli-ai
cd ameli-ai
pipenv install

How to run:

pipenv shell
python main.py

Install on windows using scoop (check website https://scoop.sh):

(inside ameli-ai's folder)
Install on windows using scoop (check website https://scoop.sh):
(inside ameli-ai's folder)
scoop install git
scoop bucket add main
scoop install pyenv
pyenv install 3.10.5
pyenv global 3.10.5
python -m pip install pipenv
scoop reshim
pipenv install

Finally, to run:

pipenv shell
python main.py

**BEWARE WITH BUGS. YES, YOU'LL FIND A LOT OF THEM **

Roadmap

• GUI available in KivyMD - WIP
• Localization support (Weblate) - WIP
• TickTick and TodoIst integration - WIP
• A dockerized server appliance
• MusicBrainz support to get current playing music title - WIP
• Simple integration with IBM Watson and Amazon Polly for neural voices support - WIP
• WhatsApp Integration - WIP
• Multilingual (partially for now) - WIP
• StopDepression module (first voice assistant with this feature) - WIP

Authors

• @gab-luz