Athena, which happens to be the name of my computer, is a vocal assistant project that I have been fiddling around for quite some time.
It is meant to run in the background as a daemon, so it detaches from the terminal when launched and closes the stdin, stdout and stderr streams.
You can monitor it through the log file located in temp/output.txt
It uses Snowboy as a Hotword Detection Engine, and then Google's speech recognition APIs to interpret the recorded message.
Snowboy works with trained models (.pmdl or .umdl files) that determine what keyword to respond to.
You can either use one of the universal models offered by Snowboy (included in resources/models/) or get a personal one using Snowboy's API, through a project like this.
To install:
- edit the
src/config.pyfile, adding the absolute paths to thetemp/directory,resources/directory, and the model file - in the main directory, run
pip install -e .
To start the daemon, run athena start, to stop it athena stop