Read the statement by Michael Teeuw here.
Amazon Echo/Alexa
-
Voice control will be great, I’d be super interested in this. I know there is a fork of this MagicMirror that does user voice control with googles API. The problem with this is that google only allows 50 API calls a day and there is no option to even buy more calls. As you can imagine the 50 calls get used up pretty quickly.
I’ve seen another Mirror that uses Jasper which is open source and does not have the API limitation, but I found this application difficult to install.
-
I was thinking more like Wit.ai, but that would be a little complicated to add. It requires you to make your own system, but you could automate anything you want using your mirror as the audio input source!
-
Hey, I am actually interested in adding Amazon Alexa at some point, and I found a tutorial for adding this. Once I finish some other stuff going on, I will be back to start working on combining the magic mirror with amazon Alexa. I think the most difficult thing is going to be getting Alexa to feed stuff directly into the magic mirror. I am looking to have her used to search and play podcast videos from NPR/YouTube, but I think this might be a bit too much for the mirror- at least for right now.
-
I thought I heard somewhere that the Alexa’s open source software doesn’t actually allow for voice activation, but rather a button must be used to activate voice commands… perhaps that’s incorrect. I’d love to add voice command to my mirror.
-
@patm189 maybe CMU Sphinx could be used as the always listening part and as a button for Amazon Echo.
-
@patm189 I see why that might be an inconvenience, and I feel the same. I would much rather use something similar to just saying “OK Google” or “Alexa, turn on…” to activate the service. However, if this is true then I will find a way to use the MMM-Button to activate it and have the omxplayer either activate default the first time the screen comes on after a certain time or change the number of times the button has to be pressed for each. Either way, I think any voice recognition software would be a cool addition to the mirror!
-
I have just rewritten the MMM-Button module so it can support more than one Button at a time. Thereby we can theoretically take one Button to start omxplayer and one to listen for speech.
-
I have a pi set up with the Alexa Voice services and indeed you must use a push button to activate listening mode. It does work pretty well. What I have started on is using the Rasberry Pi/Magic Mirror as a MQTT client. If you have an echo, you could call up and interact with Magic Mirror and your home automation and maybe have some fun “conversations” with your Magic Mirror. I think the mirror would make a great feedback device for some simple sensors around the home. You kind of end up with a home-brewed JIbo without the 500 dollar cost!
-
I also have Amazon alexa running on my rpi alongside the MM2. I say mirror, mine is on a monitor on my desk and I use it as dashboard/calendar/newscenter. I also have amazon alexa running on it. I used the guide amazon uploaded.
So the way I currently have it set up is I have the MM running on full screen and have put the alexa module as always “show on top” so I can use a mouse button to click and control alexa, while at the same time still see the other modules. Ideally, I would like to embed the alexa voice control module in the MM2. Is that something that would be possible to do?
-
Yesterday I found an update on amazons developer blog: https://developer.amazon.com/public/community/post/Tx1WVAG8L2OOXF0/Alexa-Voice-Service-Introduces-API-Update-that-Makes-It-Easier-to-Build-Hands-Fr
Now the wake up word (or voice activation) is provided. I will test the AVS/Alexa sample code during the next weeks: https://github.com/alexa/alexa-avs-sample-app/wiki/Raspberry-Pi
Or do I missed a „ready to use“ Alexa module for magic mirror?
kin regards, mick
PS: about limitations (of request etc.) I do not know, and I did not read the AVS documentation now.
