Voice assistant - education project for applied AI course.
Initially designed for Raspberry Pi model B (the old one) this project contains some derived limitations and hacks:
- run
install_pyaudio_with_patch.sh
withsu
rights to make pyaudio work in Raspbian OS (according to this) - not sure you can make it run in Windows...
- Copy
config.ini.template
toconfig.ini
. Tuneconfig.ini
. What is important? If you use voice, choose correct playback and recording devices. Callsounddevice.query_devices()
in python to learn their ids. For voice recognition use services, Wit and IBM are easy to setup. - Specify host IP and port if you use web version (coming soon).
- Run
install_pyaudio_with_patch.sh
with superuser to make pyaudio work in Raspbian. - Run
setup.sh
to install binary dependencies. - Install
pip3 install -r requirements.txt
to bring python3 dependencies.
Launch python3 larisa.py
. By default the system triggers to John
name. Say or type John
... wait ... stop
to finish.