Impetus

I made this library when I lost my voice over the winter in 2022. I wanted to still “talk” at home, and texting wasn’t as satisfying since it felt like a one-way conversation, so I made this!

Implementation

Goals

So, pyttsx3 already exists as a pleasant and easy to use offline (!!) text-to-speech library. However, for a layperson, there’s still a little work involved in setting it up properly, and my goal was to create something that reduces that barrier.

Design

In the end, I opted to keep the text to read in a file, such that if someone wants to use this, they just need to update that file and rerun the program and it’ll say the contents of the file.

Voice Selection

I also wanted to be able to pick a voice that felt like “me” when I “spoke” at home. To do so, I iterated over all English speaking voices until I found one that sounded natural[ish].

Results

My partner and I enjoyed “conversation” as usual. It took a little longer on my end since I had to type up what I wanted to “say” first, but we had a good time! The default voice I used had a very funny way of saying, “Yeah!” which made for a fun inside joke in the long term as well.