This version is quite outdated. Might be still useful somehow.

Current development is on GitLab: https://gitlab.com/open-assistant/oa-arch

Open Assistant is a prototype open source voice assistant able to automate operating system tasks and engage in simple scripted conversations.

This is "Version Zero" of Open Assistant, which is highly Linux dependent.

Check out what kidsodateless did with it: https://youtu.be/D99V9Ge9IaE

• Arch Linux or Ubuntu Linux
• Python 3.5
• PocketSphinx 5PreAlpha
• GStreamer-1.0
• GStreamer-1.0 Base Plugins
• Python-GObject
• Python-Requests

• aplay - console audio player
• plaympeg - console mp3 player
• projectm - visualizations responsive to sound
• wmctrl - window manager control. opening, closing, resize, switch windows.
• xdotool - command line x automation tool
• xvkbd - virtual keyboard for x

• Download and extract the latest Open-Assistant-master.zip package.
• Edit oa.sh to choose your desired variables, then save.
• Make oa.sh executable with: $chmod +x ./oa.sh
• Launch ./oa.sh
• Command variables will be set and oa.py will run.
• If $MINDDIR/words/commands.json has been changed and your machine is online, a new dictionary and language model will be created via the Sphinx Knowledge Base Tool.
• Say "Hello Stella" and listen for a reply. You will need to adjust your microphone settings and avoid noisy environments for best results.
• To learn and change commands, edit words/commands.json. Save, quit oa.py, then relaunch oa.sh to update.

OA Main Site: https://openassistant.org/wp/

OA Wiki: https://openassistant.org/wp/wiki

Maintained by working group lead by Andrew Vavrek, Open Assistant is an extension of Blather by Jezra, Kaylee by Clayton G. Hobbs, and includes work done by Jonathan Kulp.

Patreon: https://www.patreon.com/openassistant