Open Source Virtual Assistants, Part Two

Ater my previous post about open source virtual assistants, I looked at some of them, and decided what I really, really wanted to do was just start from scratch. I have no real intention to create anything with any purpose besides my own purposes. I wanted to dig in learn what I needed to create some semblance of a thing I want to see. And I think it will be fun to start from scratch.

I'm calling this project "Jane" because I wrote a book once, and that was the name of the AI/Assistant the main character had. Sort of a fantasy.

Anyway, the peices of this right now are:

  • A Raspberry Pi (I realize I may need something with more horsepower later.)
  • Speech to Text
  • Command Parser
  • Modules for extension to things like email, twitter, web, etc.
  • Text to Speech

I'm starting out with the goal of having this assistant do only 2 things to start:

  • Tell me my schedule for the day.
  • Tell me the weather.

I'll be adding more commands once I get these two done.

I'm using Python (of course.)

Where I am now:

I have a Raspberry Pi 2 B+ (I also have a Pi 3 B+ as well, which I might use too.) I have this microphone. I have a copy of the most recent version (Stretch) of Raspbian on an SD card. I've installed Python 3 and assorted things I'll need.

I'm starting here for speech to text. I'm hoping to get to use "Q", the gender neutral voice, but I don't know if that's possible.

The next post will update with the details of the speech to text, and how it worked.