Do a Conversation with youre HomeAssistant-Assist without needing a WakeWord

[placid nymph]

Do a Conversation with youre HomeAssistant-Assist without needing a WakeWord the hole time.

The result will be that you can have a much more normal conversation with the voice assistant than before when you had to use the wakeword again and again. You can activate and deactivate permanent listening. So you always have the choice between I always want the wakeword now or I only want it once because I want to have a conversation.

This is not a finished integration. The result is achieved using various scripts and helpers. I give no guarantee that this variant will work for you. Everyone is responsible for checking the code and any damage it may cause.

First of all, I would like to tell you about my setup. It may be that this also works with other systems, but I have the following setup:

An ESP32 with an INMP441 microphone
HomeAssitant software
Extended OpenAI Conversation Integration
Alexa Media Player Integration
The Home Assistant Cloud
Amazon Echo Dot for playback

Here I made a manual scripts and Codes
https://github.com/Coketrd/NoWakeWord_DoAConversation

And here a Video with the result

https://youtu.be/cqC0aW5u_oM

[wintry jay]

Ooh!! Someone asked about being able to do this during the birthday stream, I think!

[lilac dune]

@placid nymph please check this solution https://github.com/tetele/onju-voice-satellite/discussions/90

[placid nymph]

I also combined it now with the streamassist integration. There its now working too. So now you can talk/conversation with your 2way surveillance camera and control youre house

[placid nymph]

this looks like the same, with the different that the oncu gives you the end state from playing the tts-file? the echo unfortunately didnt do this. Because of that i must use the length from the mp3 file. or you know a other way. i have a few alexas in my house so want to use this further. the oncu for the google speakers is a very nice solution 🙂

can you show the hole oncu code?

[lilac dune]

The code is identical to what is in the repository https://github.com/tetele/onju-voice-satellite/blob/main/esphome/onju-voice-microwakeword.yaml

I understand your idea and it is very good, but I use Onju Voice for voice communication and I don't listen to music on them, and I also have HomePod mini speakers or LG home theater which have much better sound quality 🙂

[placid nymph]

yeah. in the next step i want to use the assistant with the fully kiosk dashboard tablet. Because in the living room this device is there, so i want to use the things i still habe in use 🙂