Hi openHAB community.
I was working the last months on adding services related to voice control.
GoogleSTT: Uses Google Cloud, 60 min/month free tier. README.md
WatsonSTT: Uses IBM Cloud, 500 min/month free tier. README.md
VoskSTT: works offline, you need to download the model for your language. README.md
I think I won’t add more of those services,
I still miss to have a good keyword spotter that do not requires license but I haven’t found anything, let me now if you known a good tool for it.
This is the one I’m working on right now, which is what motivates me to build the other ones. It’s a customizable Human Language Interpreter, which allow you to define your own templates to match command to write/read from your items. Allows to fallback to another interpreter, to use this one just for customization.
Note that the only audio source ready to use on openHAB (at least the one I was using to test all this) is the System Audio Source which you should select in the openHAB audio config. Other general configurations for this services are under the voice category.
Let me know if you are able to take advance of this services and how good they perform to you. For the HLI I will open another post when is fully ready but it’s on a functional state, you can test the examples at the end of the readme.
Hope they are useful to you, for me having a customizable voice system (on openHAB) was my new year challenge!
If you need some guide on how to setup them don’t hesitate to ask.
I have added a PR here to access to the Pulseaudio sources from openHAB,
as commented there, still need some work to make it run perfect with the dialog processor.
Also I like to add that I’m able to start the dialog processor through the console command “voice startdialog”. This was added by the user lolodomo (Github) among other commands, rule actions, rest endpoints… Thank you very much for the help!