Hey Mister, is there any roadmap you might be willing to share?
The application is looking very promising and since I'm currently out of options with my hass.io installation when it comes to connecting mic, I'm looking at VoicePanel as a viable option to be the main thing to host voice control. Hell, maybe I'll even make it to compile it myself in order to inject the snips config I need :)

The questions though, how do you see this app into the future?

• combining it into one app with Home Panel?
• enabling the user to upload custom assistant.zip?
• maybe adding some sound feedback when the wake-word was recognized?

Thanks!

Using MQTT Discovery, Home Assistant can make use of sensors with only minimal configuration needed.
If Voice Panel implements MQTT Discovery, this would greatly reduce manual configuration effort and errors, especially when multiple Voice Panels are used.

Status quo

Voice Panel can report device sensor data over MQTT (temperature, light, pressure, battery, etc.).
However, each sensor must be configured manually on the Home Assistant site, as described in README:

sensor:
 - platform: mqtt
    state_topic: "voicepanel/sensor/temperature"
    name: "Voice Panel Temperature"
    unit_of_measurement: "°C"
    value_template: '{{ value_json.value }}'

To be

Voice Panel uses MQTT Discovery as described in https://www.home-assistant.io/docs/mqtt/discovery/. No manual configuration required for most device sensors, as Home Assistant will discover them automagically.

Requesting to add a configurable audio feedback when a wake-word is detected by the assistant. It can be like a beep or a "yes" over TTS.

Reasoning: Currently there is video feedback on wake-word detection. The icon is being animated. It is not working in case the device is used while not looking at the screen. E.g. the device running Voice Panel might be installed hidden so that it is just a "smart microphone".

Thank you!

The following URLs are dead:

• README.md: link to issues in support section contains a typo.
• in app (AboutFragment.kt):
  • link to support is missing a hyphen
  • link to GitHub page is also missing a hyphen

As a user I would like to see my home automation platform dashboard in the application instead of the current voice panel screen. The user could slide between the two screens but still have the ability to use the voice panel commands. This gives the user more options for their voice panel screen.

The recent Snips SDK update now uses the term 'confidenceScore' instead of 'probability'. Home Assistant Adapted to this and the Snips Component now expects a 'confidenceScore'. Therefore Voice Panel does not work anymore with recent HA Versions.

Maybe it is enough to simply update the Snips SDK to make it work again.

Please add an autostart on boot option :)

I'm often unable to connect to my MQTT broker. I have to delete cache and application settings in order to reconnect to mosquitto.

It would be nice to support non english languages with snips. Maybe it can be implemented by offering the option of using a custom "assistant.zip" file, as alterative to the default one.

Hi,

I wonder if you could help me.

I try to setup the assistant with ssl enabled. The Home-assistant is running behind a self made ssl certificate but not self-signed.
I have a demoCA and I imported the CA certificate into the Android.

When try to connect I see in mosquitto log:
OpenSSL Error: error:14094416:SSL routines:ssl3_read_bytes:sslv3 alert certificate unknown

I am testing this app on a Samsung galaxy s8 and noticed that the app uses a lot of battery - the phone heats up andit is warm to the touch - even the fast charger cant seem to keep up with the power usage.

I had it open for 4 minutes and the Battery usage of the app is is 2.2%?
No other features are configured (camera streaming and sensors are all off)

Is snips really using so much power just to listen to the wake word?

The description in Android play store has a typo in the URL: vioce instead of voice, which causes a 404.

Request
Please make siteId parameter of the hermes/intent/ configurable via the Voice Assistant settings.

Motivation
This customization feature is required in order to distinguish intents in case there are more then one device running Voice Panel in the house. Ability to configure siteId will enable context-specific commands.

Examples
Based on siteId there might be different reaction on such a command like "hey Snips, turn on the lights".

Currently observed behavior

• I see the following in any intent received by HASS: "siteId":"default"
• Sample:
  {'topic': 'hermes/intent/hass:HassTurnOn', 'payload': '{"customData":null,"input":"turn on the flower","intent":{"confidenceScore":0.9853135,"intentName":"hass:HassTurnOn"},"sessionId":"a7326574-32c6-483e-9a68-9f392a6abadc","siteId":"default","slots":[{"confidenceScore":1.0,"entity":"hass_entity","range":{"end":18,"start":12},"rawValue":"flower","slotName":"name","value":{"value":"flower","kind":"CUSTOM"}}]}'}

References
https://docs.snips.ai/reference/dialogue#intent
siteId: String - Site where the user interaction took place.