[iminsandiego]

I finally got this going after a couple of hours and then switching to heroku. This is a cool proof of concept, but you must manually add the names of tv shows and movies to the utterances file. Have you thought of a way to automate adding these names? If we could simply add a way to search our movies or tv shows, that would be ideal.

[m0ngr31]

You don't have to add every show, but the more, the better. Most of the slows in the utterances don dike, I don't even have. It works great for me.

[iminsandiego]

I'm using the remote with my fire tv to control kodi, so most of the commands are unnecessary for me since navigation is awesome with the remote. The one major feature missing for me is voice search. I would like to say "kodi search for movie shrek" and have it open the search, automatically type shrek, and then press done.

Any ideas how we can add that ability?

[m0ngr31]

I don't know if that will be possible with the way the Kodi API works. I'll look into it though.

[iminsandiego]

Here's another Alexa-Kodi project that's very similar to what you are doing.
https://www.reddit.com/r/amazonecho/comm...ing_media/
https://github.com/rsummers618/HTPC_Alexa_Skill

I installed this one as well as yours. This one has the advantage of not requiring Heroku. It did work well for me. The features he mentioned like searching Netflix and Pulsar won't work on the Fire TV, but likely work well on some other platforms. He requires chrome launcher in order to call Netflix and that's not calling Chrome properly on FireTV.

[Gyroscope352]

Hey! Just discovered this yesterday and really excited to try it out.

Is OpenShift still having problems? I'm getting the "There was a problem with the requested skill response" error when telling my Echo to "ask kodi", and "The remote endpoint could not be called, or the response it returned was invalid." error when testing on Amazon's web site. But it sounds like people are having trouble with Heroku as well? Which one is the more reliable choice at the moment?

That other one posted by iminsandiego looks nice, but doesn't seem to be as powerful as m0ngr31's, at least in terms of the stuff I want my Echo to do.

[iminsandiego]

Hey! Just discovered this yesterday and really excited to try it out.

Is OpenShift still having problems? I'm getting the "There was a problem with the requested skill response" error when telling my Echo to "ask kodi", and "The remote endpoint could not be called, or the response it returned was invalid." error when testing on Amazon's web site. But it sounds like people are having trouble with Heroku as well? Which one is the more reliable choice at the moment?

That other one posted by iminsandiego looks nice, but doesn't seem to be as powerful as m0ngr31's, at least in terms of the stuff I want my Echo to do.

I didn't have any luck with OpenShift after a couple hours. Heroku was a little easier but you have to adapt m0ngr31's instructions a bit to get Heroku to work.

[Gyroscope352]

I didn't have any luck with OpenShift after a couple hours. Heroku was a little easier but you have to adapt m0ngr31's instructions a bit to get Heroku to work.

Thanks!! I was a little confused at first but somehow managed to stumble my way through Heroku in just an hour or so and I have it running perfectly! This is so so so cool. THIS is the power of the Amazon Echo. I'm incredibly impressed. Hopefully I don't run into too many snags in the next few days (m0ngr31 has me all nervous about Heroku's limits, but we'll see how it works!)

Thanks for putting all this together, m0ngr31. You rock.

[m0ngr31]

I updated the documentation, hopefully it's a little more clear now. I've also switched all the code/documentation to run off of Heroku now instead of Openshift.

[geman220]

Just wanted to report that I've been running flawlessly with OpenShift for about a month now. The only minor issue is you have to sort of "wake up" the service. If I haven't asked Alexa to play a movie recently the first time I ask "Alexa ask TV to play the movie Star Wars" it will respond "the requested skill took to long to respond". After that I can ask again and it will work. It will continue working after that until you don't make any requests for several hours. I could not get it to work at all on Heroku in any situation.

Edit:
To clarify things like "pause" "play" etc would work. But asking to play a show/movie or asking if there are new shows/movies would not work on Heroku.

[edru]

nm-found post that addressed my issue.

[edru]

I was able to get this working through my home server. If anyone is looking to modify the readme for implementation on their own server, here are the basic requirements.

Add mod_wsgi to apache with apt, generate a self signed cert with this guide.

In the wsgi.py script, you'll need to comment out (or remove) lines 4-9, as they are specific to Openshift deployments.

In your default-ssl.conf file for your SSL apache configuration you need to add a line for WSGIScriptAlias, something like:

Code:

WSGIScriptAlias /kodi-alexa /var/www/html/kodi-alexa/wsgi.py

In Alexa Skill Set, Add your IP and service name as the HTTPS URL: https://XXXX.ddns.net/kodi-alexa
On the SSL Certificate page, copy and paste the contents of your apache.crt file generated from the guide above.

Finally, if you haven't already, you'll have to open your apache server to the Internet with port forwarding.

All in all it works pretty well. If I ask for playing a random movie or episode, occasionally I'll get Alexa telling me that the "Requested Skill took too long to respond," but it will work. Probably a timeout that I need to find and change.

HTH, available to answer questions if I can.

Thank you for this. This is exactly what I needed to get it to work.

[kevcompton]

please ignore

[edru]

Loving this setup. Favorite thing to do is to ask it to play a random show. Thanks to MakerMusings and yourself this is getting along nice.

Just a few command requests:

Clean Library
Update Library
Play newest episode of SHOW
Play random unseen episode of SHOW

[m0ngr31]

I just pushed a pretty big update that should parse phrases better. I also updated to include those commands edru.