@darrene You’re missing parenthesis in your call to frombase64:

this.frombase64();

Atom based mini PC should be enough. Something like this:

https://www.amazon.com/MINIX-NEO-Z64-W-Directly-Technology/dp/B016OBOJ5U/ref=sr_1_31?s=electronics&ie=UTF8&qid=1477430450&sr=1-31&keywords=mini+pc

@bhepler you’re right. putting a delay should help with the CPU usage.

What would be the point in power saving if you don’t turn off the HDMI?

@jeremytripp So the idea behind the voice control module is that it’s responsible for recognition of the command and sending a global notification to any other module interested in this.

As for the module that you’re describing like displaying some text based on the command I would recommend creating a separate module in which to listen to this specific message and doing what you need it do.

@jeremytripp So, just to confirm you’d like to activate/deactivate the mirror with voice commands without walking into the room, right? I don’t see anything glaringly wrong with the code except for potential conflicts such as the scenario when you’re in the room where the mirror is and you say “Go to sleep”, the monitor will be switched off, but when you start moving it will switch on again via the motion detector.

I find the sound quality of my USB cam for voice recording quite usable. What’s more important for the snowboy recognition to be more accurate is to record your own voice for the model on the same device you will be using it after.

To implement globe module should be pretty straight forward. However RPI wouldn’t have enough power to handle smooth animations with WebGL, hence the choice to use Mac Mini for this project.

@James thank you for reporting the bugs. I will try to fix them soon.

The code doesn’t save the photo into the file. The image data base64 encoded is passed from the camera module to the node_helper. I could add this, but am I afraid of filling up the space on the RPI.

Which provider did you configure the nodemailer with?

It looks like you’re using google speech APIs. You are quickly going to run out of the quota and start receiving error messages.

Just to connect things, here’s the module:

https://forum.magicmirror.builders/topic/603/camera-module

@KirAsh4 sure. Cameras for everybody!

I’ve just published the camera module. Besides displaying a video feed on your mirror, it could also take photos and email them.

[card:alexyak/camera]

@KirAsh4 sure :) Feel free to branch it and add this functionality. I am done publishing it, btw. I will make a post on the modules topic.

I’ve got a camera module working on my mirror. It works together with my Voice Control module and is able to make selfies and email them to a specified email address. I just need to find some time to package it and post it on github.

@Mitchfarino Sorry, I’ve not tried this yet myself, but have you tried to use this option that you have selected: “From Existing Node.js code”?

@Mitchfarino Most certainly you should be able to. There are node.js specific projects. You just need to install tools:
https://beta.visualstudio.com/vs/node-js/

[card:alexyak/voicecontrol]

I’ve updated the VoiceControl module to display a list of commands on the mirror:

[ card : alexyak/voicecontrol ]

thank you, @strawberry-3.141. But couldn’t we use the git url as dependencies in the package.json?

Unrelated questions how do you insert github-card into a post?