Amazon Alexa TTS from Pimatic

---

Hello everyone.

So I have been searching for this for a year now, and looks like finally someone have been able to make tts to Alexa without having to ask her for the latest notifications etc…

It is made for home assistant, but I’m hoping we have some clever developer in the pimatic forum who can port this to all of us to use.

Watch the video which explains a little about it here
https://youtu.be/Q2CHrOK2DbM

I’m sadly no developer, so I’m not able to port this myself, but looks like the ground work has been done now.

What is your thoughts?

More information
https://community.home-assistant.io/t/echo-devices-alexa-as-media-player-testers-needed-it/58639/8
https://github.com/keatontaylor/custom_components/tree/master/media_player

It is made in node.js like pimatic, so I guess it is should be able to be ported to our platform?

Am I really the only one who is interested in having Alexa talk to me?

Examples:
Someone is at the front door.
Or
Your laundry is finished, remember to take it out of the washer.
Or
You have received some mail in the your mailbox.

There are endless possibilities if we can send some tts commands directly to Alexa.

Hi there. As we are in the middle of the holiday season it may take some time before someone picks up on this.

As far as I understand this, an Alexa TTS feature requires a specific Alexa skill to be written. Anyone who he is interested to look into this may also want to look at the following Alexa example which demonstrates the use of Alexa Audio Player for skills.

https://github.com/alexa/skill-sample-nodejs-audio-player

After reading the thread mentioned in post #2, it seems, that for just using the TTS feature, an Alexa skill isn’t required. It seems much easier. Actually i’ve already triggered a notification on Alexa by a Pimatic rule This is the interesting post:

https://community.home-assistant.io/t/echo-devices-alexa-as-media-player-testers-needed/58639/4

The code starts an webserver, which is waiting for http post requests to trigger a TTS notification on Alexa. So it’s very easy to add an Alexa notification to Pimatic. For example:

execute "curl -X POST -d \"tts=Das ist ein Test\" -d \"deviceSerialNumber=G090LF096457XXXX\" http://192.168.1.2:8091/alexa-tts"

Here are some hurdles:

• Since the code uses Nightmare to login into the Amazon Webpage, a running X-Server is required. My Pimatic Raspi is headless and there’s no X-Server running. I’ve experimented with Xvfb, but wasn’t successfull. Therefore i’m running the code on my Windows 10 PC right now.

• Additionally to USER and PASS, the environment variable URL has to be set according to your location (e.g. alexa.amazon.com for US users). Here in Germany it’s “export URL=https://alexa.amazon.de”.

• 2-Factor Authentication has to be disabled for Amazon login

• After starting the code, you should get a “Logged in” message in the console. At first that wasn’t the case for me. It helped a lot to change the “show” Parameter of Nightmare to “true”.
  var nightmare = Nightmare({ show: true }) at alexa-api.js#3
  This opens a browser window, where you can check the login process. In my case an additional authentication step was requested in the window (entering code sent by email). Glady this had to be done only once.

• You can get the deviceSerialNumber, which is required in the curl command, from the webserver:
  curl -X POST http://192.168.1.2:8091/alexa-getDevices

Maybe someone else wants to tinker with it.

Since there are too many hurdles in the solution in my last post (especially the required X-Server), i was looking for a simpler solution. And i found one, which is dead simple It took me less than 10 minutes to set it up. It’s a bash script called “alexa_remote_control.sh”. You can get it here:

https://github.com/thorsten-gehrig/alexa-remote-control

I just changed username/password in the file and ran it. Login to Amazon was successfull at the first time. Then i’ve added this rule to Pimatic:

execute "/home/pi/scripts/alexa_remote_control.sh -d Schlafzimmer -e speak:'Das ist noch ein Test' > /dev/null 2>&1"

Et voila. Alexa was talking to me

The script is configured for German users. For other locations you have to change some variables at the beginning of the file, which are commented out.

More info (in German):
https://blog.loetzimmer.de/2017/10/amazon-alexa-hort-auf-die-shell-echo.html

@googlyeyz
Very nice. I will definitely try and get this working, if I’m able to. Would be so nice to have Alexa talk to me when something requires my attention instead of me having to ask what my notifications are. Thank you for the digging, I hope I can get it working with your info provided.

Thank you.

Kind regards Simon

@simonbove

I was looking for TTS from Pimatic to my Echo Dots, when i bought them 1,5 years ago. But it wasn‘t possible then. Thanks to your post, i discovered, that TTS is working now. Btw. besides TTS, the script can trigger many other Alexa Actions by Pimatic, e.g. play radio station, listen to flash briefing, start a routine, …

@googlyeyz Hi.

I’m on my phone (don’t have a computer with internet for a while) so that might be the reason, but I can’t for the love of god find where I need to put my login credentials in the the Alexa-remote-control.sh file?

Could you point me in the right direction, since you got it working already?
Can’t find anything about changing the location either, I’m using the .co.uk domain.

Thanks.

@simonbove said in Amazon Alexa TTS from Pimatic:

Could you point me in the right direction, since you got it working already?

line 48-49:
EMAIL='amazon_account@email.address'
PASSWORD='Very_Secret_Amazon_Account_Password'

Can’t find anything about changing the location either, I’m using the .co.uk domain.

The script doesn’t contain the addresses of the UK server. Only German and US servers. I don’t know, if somebody has successfully used it on the UK server. But maybe these changes are working:

LANGUAGE="en-GB"
#LANGUAGE="de-DE"
#LANGUAGE="en-US"

AMAZON='amazon.co.uk'
#AMAZON='amazon.de'
#AMAZON='amazon.com'

ALEXA='layla.amazon.com'
#ALEXA='layla.amazon.de'
#ALEXA='pitangui.amazon.com'

I can confirm, that it’s working on the German server. But only if 2 Factor Authentication is disabled! According to an issue on GitHub, there seems to be a login problem on the US server. Apparently Amazon has tightened the security for US users and presents a captcha, which has to be solved in a browser, on each login. This makes the script unusable in this case. Hopefully this doesn’t happen on the UK server. Just give it a try.

https://github.com/thorsten-gehrig/alexa-remote-control/issues/10

Thanks for the reply.

I have tried to change the script to UK as you wrote, no errors when I’m running the speak command, but I’m not home so I don’t know if it is working or not in real life. Won’t be able to check before 1,5 months from now. So patience is needed hehe.

If you don’t want to install jq for using tts, you can also use
alexa_remote_control_plain.sh
from the github repo, that works the same way for tts, but a few other functions will not work.

@GooglyEyz how is it about security? Do you know a way to not write the password in clear text in the script?

@tech_diem said in Amazon Alexa TTS from Pimatic:

@GooglyEyz how is it about security? Do you know a way to not write the password in clear text in the script?

No. Actually i‘ve stopped using the script, because besides clear text password, it doesn‘t support 2 Factor Authentication. I don’t want to disable 2FA on my main Amazon account. Perhaps creating a 2nd account without 2FA and use it only for Alexa would be a solution. But then i‘d have to migrate all my Alexa custom skills to the new account.

So I got it working finally.

I used the alexa-remote-control_plain.sh script from above mentioned link.

wget https://loetzimmer.de/patches/alexa_remote_control_plain.sh ./alexa_remote_control_plain.sh

Edited the script for the UK Alexa using below:

sudo nano ./alexa-remote-control_plain.sh

LANGUAGE=“en-UK”
AMAZON=‘amazon.co.uk’
ALEXA=‘alexa.amazon.co.uk

And your Amazon account email and password

Then I

sudo chmod +x ./alexa-remote-control_plain.sh

Then you should be able to run:

sudo ./alexa-remote-control_plain.sh -a

Which should return your echo devices

When you have found the name of your echo device, you can use the tts with below

sudo ./alexa-remote-control_plain.sh -d YOUR_ECHO_DEVICE_NAME -e speak:'I can speak'