New add-on: OpenAI API speech-to-text for voice assistants

Share your Projects!
,
1

Hi all, I’ve built a new add-on that’s meant to be an alternative speech-to-text provider for voice assistants.

What it does

For those using a voice assistant, this add-on utilizes the OpenAI API to process voice commands, so you don’t need to manage any speech-to-text models locally.

It takes about 1.5 seconds to process a 4-second voice command, and my level of usage costs me about US$0.02 per month in API costs.

Why I built it

I bought a Home Assistant Voice and set up a local speech-to-text engine (Whisper), but as noted in the docs, processing voice commands took up to 8 seconds on my hardware.

I decided to try my hand at building my own Wyoming speech-to-text add-on, and here we are!

How to learn more

  1. The add-on repository is available here on GitHub if you’d like to give it a try. Setup instructions are in the add-on README.
  2. The source code used for the add-on is in this GitHub repository.
  3. I’ve also written a post about creating this add-on on my blog.

I hope some of you will find this useful! Please don’t hesitate to leave feedback, whether you liked it or not. Feel free to open issues in the GitHub repositories if you have questions or run into any problems. Thanks for reading!

2 Likes
2

hi, thanks for your work. I’m about to try some STT, ideally using gpt-4o-mini-transcribe

But any any idea what are differences between this add-on and existing

?

Thanks.

3

There’s als this one, just to keep the list complete. :wink:

4

Hi both, thanks for sharing these other projects. Looking through them, it would seem that the idea is pretty much the same, and I am not doing anything particularly different.

But feel free to let me know if there’s anything I can do to help you make better use of my project!