Unable To Link Alexa Skill

[alienstar-uk]

Scrypted Server is installed and working and connection to the Scrptted Cloud and Hikvision cameras. Whenever I try to enable the Amazon Alexa skill it fails with the message

"
Unable to link the skill at this time
Please try again later.
"

As I use a UK Amazon account and Alexa devices I use the following URL to enable the skill:

https://www.amazon.co.uk/ClockworkMod-Scrypted-Home-Automation/dp/B09WHHD9VN/ref=sr_1_1?keywords=scrypted&qid=1706554229&s=digital-skills&sr=1-1

After a brief period of trying to link the skill it fails with the "Unable to link the skill at this time" error message.

In the console (with debug enabled on the Alexa plugin) it displays the following:

"
########################
2/5/2024, 10:21:50 AM
########################
received directive from alexa Alexa.Authorization/AcceptGrant {
directive: {
header: {
namespace: 'Alexa.Authorization',
name: 'AcceptGrant',
messageId: '13439667-3d29-4426-b3c9-46e801e9c4f2',
payloadVersion: '3'
},
payload: { grant: [Object], grantee: [Object] }
}
}
response to alexa directive 200 {
event: {
header: {
namespace: 'Alexa.Authorization',
name: 'AcceptGrant.Response',
messageId: 'bbc56acb-f031-453d-b7d4-41d2a0c76e93',
payloadVersion: '3'
},
payload: {}
}
}
"

I have attempted enabling the skill multiple times over the past 5 days from an Edge browser on Windows 10 and on IOS 17.3, all wih the same failure.

Desktop:

• Windows 10
• Edge Browser Version 121.0.2277.98 (Official build) (64-bit)

Smartphone

• Device: iPhone 15 Pro Max
• OS: IOS 17.3
• Edge Browser Version 121.2277.99

Scrypted Version Version 0.91.6 running on QNAP Docker Debian Image

[koush]

try using the mobile app

[alienstar-uk]

Tried the mobile app but this also fails.

[alienstar-uk]

I recreated the QNAP container and changed the container network mode to be Custom - Host (as per your QNAP Container install instructions) rather than a virtual IP address (which I had used to prevent port clashes) and this worked.

[blairscarlett]

I am having a similiar problem. I have tried from both the mobile app and several computers. The only difference architecture difference is that Scrypted is running on Ubuntu within HyperV. I have confirmed that Scrypted Cloud port forwarding is working. I had this working on a previous HyperV yet have to rebuild it and cannot get the Alexa skill to link.

[foxy82]

I'm also suffering the same thing - using Proxmox LXC container - scrypted is saying the port forward is working and checking locally in my network I can see the container has the port open.

I see this in the cloud log just before the Alexa App/Alexa Website says linking failed:
50.219.188.186 /web/component/alexa/endpoint

The Alexa plugin doesn't add anything during linking but it does show this :

plugin loaded
startup failed Error: 'tokenInfo' is undefined
    at h.getAccessToken (/src/main.ts:418:19)
    at h.this [as syncEndpoints] (/src/main.ts:332:40)
    at h.start (/src/main.ts:99:17)

[skynet01]

I've been having the same issue for month now, it used to work ok for me. You did get further than I did by actually getting an error shown. Looks like a bug, maybe @koush can trouble shoot here? I am on amazon US server and tried desktop and app version and get same "please try again" error

[blairscarlett]

I just tried linking the skill and it worked today. I didnt change anything on my config so I have to assume it was Amazon. Also worthwhil eto mention, I use amazon.ca Glad it works now!

[skynet01]

Still the same issue for me on the US site. Port forwarding works and I can access my instance from the cloud.

[foxy82]

I downgraded to an older version that worked for me in the past. I don't see the token error now but in the Alexa plugin console logs messages I see the Alexa messages arrive after Amazon has given up trying to link the skill - so I wonder if it is just too slow / too much load on the scrypted cloud plugin.

[skynet01]

What version did you downgrade to? Mine still has an error

unhandledRejection Error: 'tokenInfo' is undefined
    at h.getAccessToken (/src/main.ts:418:19)
    at h.addAccessToken (/src/main.ts:213:40)
    at h.deviceListen (/src/main.ts:207:27)
    at processTicksAndRejections (node:internal/process/task_queues:95:5)

[InnovativeDevelopers]

Same issue here. Has anyone in the UK managed to enable this?

[InnovativeDevelopers]

Fixed it finally! There's an ongoing issue with docker on Synology where networking starts to run slow. I manually set the DNS to Cloudflare (anything will do) and everything started to work. Weird!

[alienstar-uk]

I am in the UK and have got it to work using a QNAP container. After initial problems I recreated the QNAP container and changed the container network mode to be Custom - Host (as per the QNAP Container install instructions) rather than a virtual IP address (which I had used to prevent port clashes) and this worked. Have been running it like this since Feb 18th and have had no problems. The camera feed is only about 2 seconds behind the live feed - which is far better than the MONOCLE solution I used previously where it could be up to 10 seconds behind and frequently suffered from "buffering" messages.