Actionable Notifications via Alexa Media Player

some positive side effect when writing in a different language, you make somebody laugh without knowing (or angry… it depends, iam glad that i wrote two sentences). :stuck_out_tongue_winking_eye:

I had some Ideas to try:

  • in the developer console under “Distribution” change some Description and check if this is updated (you have to hit the “Save and continue” Button below and Re-Build)
  • you could also check the JSON Editor (last item under Invocation, Intents, Slot Types) if the invocationName is here changed.
  • and you could check the language from the skill matches your alexa’s

hope that helps isolate the problem

I am assuming to update to the new release we just need to update the code on the alexa skill? Sorry I didn’t see this covered on the wiki?

at first I couldn’t link the skill either. my problem was this:

6. Under Client ID provide:

* `https://pitangui.amazon.com/`  if you are in US
* `https://layla.amazon.com/`  if you are in EU
* `https://layla.amazon.co.uk/`  if you are in the UK (You may still need to use the EU URL if things do not work)
* `https://alexa.amazon.co.jp/`  if you are in JP and AU

even though Im not in the US, for me it worked when i selected US.

I have no problem using audio tags to play mp3 files with Alexa Actionable notifications.

When testing I found that the text submitted to script.activate_alexa_actionable_notification should not be surrounded by quotes. In other words,

text: <audio src=‘https://etc.mp3’/>

works, but fails if surrounded by quotes.

My endpoints are on S3. Further errors will occur if the mp3 was not converted as described in Speech Synthesis Markup Language (SSML) Reference | Alexa Skills Kit.

You can create a second implementation of the skill and remove the .ask(‘’) from the return of class LaunchRequestHandler. The second skill just plays the text it gets from HA without waiting for a reply. Clone script.activate_alexa_actionable_notification, change the script name, and put in the new skill ID. Now you can play any SSML (of course, I haven’t tested every possible tag).

Let me know of problems or better solutions.

1 Like

this could be a better tts solution cause announce always have this beep and tts itself has no natural speak output (like pause between words).

I hope you’ve got this fixed but in case you don’t here is what I did to fix it. Go to the console in the Amazon developer area and look at the Intents section for the skill. There is probably the Hello World intent leftover from the template. Delete it. Make sure the Yes and No intents are in the list. Go to each and add at least one sample utterance to each Yes and No will work. Save and rebuild and that should do it.

Frank

I am in the US and used that setting. Still not working.

I guess I’ll delete it one more time and start from scratch using the suggestions above from @florian.ec and @FrankChildress.

another possible source of error may be Account linking, did you use it? Try long-lived token instead.

Thanks for the advice, I have changed those things and will see what happens at 1930…

I deleted everything and started from scratch once more and I think it might work. Again I’m not at home to try it but the HA sign in worked and now the skill on the Alexa app is showing “custom actions” now instead of “change me”.

So it’s step in the right direction.

The only thing I did differently is removed the “Hello World” intent, added some sample utterances and I hit save a couple of times before building and deploying the skill.

I’ll see if it works when I’m home in a few hours.

glad to hear that! Today i wanted to try a few other things and had also trouble to set it up. I see that there are a lot of changes since i did it a couple days ago. As you did, the Hello World intent was removed.

This and other intents are now easier to set up directly in the JSON Editor with https://github.com/keatontaylor/alexa-actions/blob/9c6cd64828f45a72171ebf5e760be18460a13e12/skill-manifests/locale_en_us.json

Anyway, i had trouble with the latest lambda_function.py version 0.7.0 and now use 0.6.0 for the moment. (I had some weird issue where the input_text was stuck with the text first invoked with the skill. Even the input_text changes in home assistant the skill was using the first one.)

Well, it still doesn’t work. :confused:

the reply after saying “Alexa, open custom actions” was “there was a problem with the selected skill response”.

Any other ideas?

Seems like the only time I got that I had the endpoint screwed up … but I’m not 100% sure. May have been when I had capital C in custom actions.

No real new ideas here.

Frank

did you use the correct skill-id in the activate_alexa_actionable_notification script? i got the same response this morning until i realized i had an old skill-id .

TBH, I had an old skill id the first time I tried it this morning after re-creating everything. But I got the same error again.

So, I decided one more time to try it again.

This time I realized that the procedure had changed (or I missed it before) and I needed to add the requirements.txt file contents.

And now, finally, it’s all working! :tada:

It only took 5 tries. :slightly_smiling_face:

believe me or not, i was just about to suggest this, i forgot that too the first time :rofl:
happy it’s working for you!

1 Like

Hi everybody,

Yesterday it was working flawessly, but today I updated HA and Alexa is just saying: “sorry, I had trouble doing what you asked, please try again” in a loop if I ask her to open custon actions or if a notification is triggered.

Someone knows what’s going on?

try to update the alexa skill with the latest code from github. There is some advanced error handling now and you maybe get a hint what’s wrong.

I updated and now it’s working. Thank’s.

I have exactly the same problem Alexa stays speechless
I have tried everything I can , I can trigger it from Amazon skills setup but I only get the response to my PC speakers but not Alexa,
Question
In the script , do you put the Alexa device you want to use ; or you put the alexa_device and then you specify the device in the automation
It is driving me crazy for 5 DAYS now

Any help please