Willow - Open Source Echo/Google Home Quality Speech Hardware for $50

[kristiankielhofner]

@rlkoshak

All I can say is OH isn’t HA and we don’t have the advantage of being backed by a for profit company with roadmaps and top down directed development.

I completely understand and respect that. After all Willow isn’t exactly a money printing machine either (total income: $0). I only mention HA because as seen from this discussion the user takeaway from the Willow and OpenHAB experience is “Willow can’t do that”. We can and do support “that” but of course we’re inherently limited by the communication with the platform on the other side.

This is in no way intended to be critical of OpenHAB - I’ve had open source projects for 20 years. If anyone gets it, it’s me. I just want to be clear that if/when OH does support any of this it’s likely a few lines of code in our OpenHAB endpoint for us to support it.

There are plenty of other things we can do that HA doesn’t support either. For example, we’re working on a hybrid fuzzy and semantic search autocorrection implementation for voice commands. Nothing else out there in the space has a remote concept of this so we’re having to work around it all over the place. The good news for OH users is we settled on a mechanism to support this that is common-denominator and will work across all supported command endpoints. Anyway, just an example - that’s APIs for you.

@rlkoshak

What OH lacks right now is:

1. a volunteer to make changes to that REST API endpoint along with the text interpreters to support it
2. an text interpreters that support meaningful responses in the first place
3. anything in the architecture that would allow an end user to customize what would be returned by that REST API endpoint assuming we had 2
4. a UI to support the customization of what is returned by that end point
5. a volunteer to develop an add-on to implement a more native interaction between OH and Willow

We’re a tiny team (three people part time) and essentially brand new open source project - this reads exactly like what I write almost every day when our users ask for stuff that’s way out there . The problem with giving people everything is even when you do that they just want more.

[rlkoshak (Rich Koshak)]

Don’t get me wrong. I’m not complaining. The list above is a call to action. There are several developers on this thread.

[obones]

@rlkoshak

a volunteer to develop an add-on to implement a more native interaction between OH and Willow

Creating and addon to listen for a JSON to be posted and replying with a text to be spoken out loud isn’t that hard, but that’s what in between that I feel is a bit tricky.

I mean, let’s say the addon receives this:

{ "language": "fr", "text": "Quelle heure est-il ?"}

How does it determines that the valid reply is Il est quinze heures quarante cinq ?
Does it use a dictionary of possible commands?
Does it use a language model like the one from Mistral AI ? If yes, how is this LLM plugged into openHAB ?

So many questions…

[rlkoshak (Rich Koshak)]

I don’t think it’s going to be easy by any means. But I think responses, at least to start, can be pretty stock and predetermined. If told to turn on the a light, respond with “light X turned ON” or even be more generic with “Completed”.

An add-on to OH is not intended to be a replacement for nor a whole implementation of a smart assistant. I wouldn’t expect a query like the above to even make it to OH in the first place and if it did, I wouldn’t expect OH to have a meaningful response to it. If I ask Alexa or GA what time it is they respond on their own.