MMM-AssistantMk2

A Former User

@DopeEthiop
Your issue always and only happens when you say “jarvis”? or all the same with other hotwords?
Sometimes the time lag could happen, but your case seems something wrong. Leave your issue to git repository issue board to look inside more deeply. here is not so good place to track issue.

dxfan227

Hello

I have both mmm assistantmk2 and hot word up and running however when I saw jarvis the first time it works…but then when I try again…no hot word picked up. any ideas?

bibi

Hello all, was just wondering if and how difficult it would be possible to add a kinda face animation in Mk2? I have seen Hello-Lucy and i found the idea cool but the face is kind of creepy :grinning_face_with_smiling_eyes:

Mykle1

@bibi

Take a look at this module. It displays animated faces of sorts.

https://forum.magicmirror.builders/topic/9963/mmm-ai.

Or, if you have an animated gif that suits you, you can use just about any image module to display it.

Davido

@dxfan227

have you found a solution?

if not, please attach your config

dxfan227

@Davido

Thanks, i did find a solution. For some reason the HOTWORD RESUME notification was not working. I had to “hard code” it into the .js file itself for the module. It is now working.

dxfan227

Is there a way to unlink this from my personal google account info?

Ie when I ask to show me photos of dogs…it actually goes through MY google photos account and shows photos of my dog.

Cr4z33

@Sean can you tell us please anything about what’s coming with versione 3? :smiling_face_with_halo:

A Former User

@Cr4z33
Well, now it is on development and not ready to release so details could be changed.
But it might have these features;

No need to wait after hotword. You could say “mirror, blahblah” instead “mirror (pause and wait activation) blahblah.”
Some annoying dependency would be removed. HTML5 default audio output would be used, so you don’t be bothered by sound output configuration. If your web browser can play some audio, it would be enough
Installer will be provided.
Recorded wav file by user or other module could be used for your query. (So User can query via 3 ways - text, real-time vocal query, pre-recorded vocal query)
Annoying center region issue would be gone.
ScreenOutput would be more controllable (duration time, size, activation inner links, suggested commands, …)
- The core informations of screenOutput would be extracted for later using. (e.g: external link to open a broswer for getting additional information. it could be transferable to other module for later using.)
Creating and managing custom action would be easier.
- So I’m considering to remove transcriptionHook trick to use action instead.
- I served transcriptionHook for making simple voice command, but that is not a natural way to use Google Assistant. In many cases, Ppl tried to make commands which are not related with google assistant directly. Rather those kinds of commands could be achieved by just simple voice detection without AI. So I now add that feature to MMM-Hotword and being condering to drop out from MMM-AssistantMk2
YouTube features will be removed. (And Spotify also)
- Instead, YouTube, Spotify, Web Browsing, Map Direcions, Image search would be separated as independent module. those modules will use the info which is extracted from result of query. Of course, each module could be used standalone without AMK2. I’ll make it one-by-one. (Or whoever has interests could)

Cr4z33

@Sean wow that’s a BIG change in the module! 😲

Can’t wait to start trying it out mate to use all the good stuff you are adding in it.

Thanks for letting us know and have a nice Sunday. 😁

saljoke

Wonderful. I want to try these changes!!!

smarthome

Hello
How can I stop Assistant’s response and make request again like Google Home Speaker can ?

smarthome

This is my config, still don’t work yet

{
    module: "MMM-AssistantMk2",
    .....

      notifications: {
    ...
        ASSISTANT_UNDERSTOOD: "HOTWORD_RESUME",
      },
    }
  },
  {
  module: "MMM-Hotword",
  ...
    models: [
     ...
      {
        hotwords    : "stop",
        file        : "stop.pmdl",
        sensitivity : "0.5",
      },
    ],
    command: {
     ...
      "stop":{
      notificationExec: {
        notification: "ASSISTANT_CLEAR",
        payload: (detected, afterRecord) => {
          return {profile:"default"}
        }
      },
      afterRecordLimit:0,
      restart:false,
    },
  },
},

A Former User

@smarthome Sorry, at this moment, there is no way to interrupt response. Just wait ends.

smarthome

@Sean said in MMM-AssistantMk2:

@smarthome Sorry, at this moment, there is no way to interrupt response. Just wait ends.

Tks for your quick answer! Can we make a command exec to mpg21 to stop blayback google response?

smarthome

And other question is, if I ask a command that Google can not find the suitable answer, there’s no response while google home speaker give a promise answer? Is there any configure to do that?

A Former User

@smarthome

Can we make a command exec to mpg21 to stop blayback google response?

Ideally, might be possible. But not only playing sound is the issue related. So, currently not so easy.

And other question is, if I ask a command that Google can not find the suitable answer, there’s no response while google home speaker give a promise answer? Is there any configure to do that?

Currently, No.

Anyway, I’m building a new updates. It might have possibility of solving your issue. I’m planning to remove dependencies about sound out. It might become more controllable about voice response.
But due to my vacation, and new RPI4, Raspiban Buster and other many things which I should concern besides make its release somewhat lately. sorry.

smarthome

This post is deleted!

smarthome

@Sean
How can I remove “Repeat after me” whenever ask ASSISTANT _SAY to speech text?

A Former User

@smarthome Impossible. ASSISTANT_SAY is not perfect solution to make mirror saying something. At any time it could be changed by Google how it works. When I implement that feature at first time, there is no additional “Repeat after me” spoken phrase, but Google changed it’s working way. So I didn’t recommend using it. Use real TTS solution to make your mirror say something.