question

Alexander S Polson avatar image
Alexander S Polson asked

Creating a smoother experience for long-form content?

I'm in the process of developing a skill that is responsible for reciting the contents of articles back to the customer. Sometimes these articles can be rather long, 20-30 pages or more. I'm trying to come up with an effective means of interacting with the customer such that the reading progress is regularly saved, but also keep the requirement to directly interact to keep reading to a minimum. Unfortunately, I've only been able to manage the first part successfully.

Currently, after each page is read, I tell the customer how far through the article they are and ask if they want to keep reading or not? If they say yes, then the next page is read. This continues until they get to the end of the article, at which point the next one is presented, or they quit the skill.

What I'd love to be able to do is to eliminate the need for the customer to say they want to continue reading the article and instead just have the next page read without any action taken, while also saving the gradual progress being made. I would love it if, after speech has been rendered on the device, a subsequent call could be kicked off to my skill to return the next page. I've poured over the ASK documentation and can't find anything that suggests interaction like this is supported. Perhaps there's another way that I'm not considering that would potentially make this possible?

how-tointeraction
10 |5000

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Rokas avatar image
Rokas answered

This is impossible to do without those interactions, because of the limited amount of text that can fit in ssml.

The only way that it could be possible is to use polly to read the text and build audio file from it in the AWS that you could play using audio interface, but this is a very hacky solution and I am not sure if that would work.

I suggest to post your feature request in alexa uservoice here, so that alexa team would know what is important for the developers.

10 |5000

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Alexander S Polson avatar image
Alexander S Polson answered

Thank you for your response RokasV. It's been a while but thought you should know that this is the route that I ended up taking along with making mine a music skill so the bits of audio can be queued up relatively transparently to the customer. It actually works pretty well, though the setup isn't as optimal as I would like, and it doesn't use the same voice as is used for non-Polly-based skills (not a huge deal on that one).


I posted a feature request as suggested but alas it's not gotten any attention.

10 |5000

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.