question

Sanjay Chadha avatar image
Sanjay Chadha asked

AWS as a Speech to Text Engine

Few questions regarding Alexa/AWS: 1. Can we replace keyword Alexa with our keyword? 2. Is the keyword locally matched to know keyword/prompt was heard or is it sent to the server to check. We find it is slow if they keyword is also sent to the server. Where is prompt recognition done in Echo? 3. We would like to use AWS as a Speech to text engine i.e. when we send the speech to AWS we need to receive back text as opposed to sound file as a reply - (similar to Google Speech API). Thank you for you time.
alexa voice service
10 |5000

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

Eric@Amazon avatar image
Eric@Amazon answered
You must use the Alexa keyword if you expect to commercialize your product and have it be certified by Amazon. However, we do not currently provide a wake word solution, so there is nothing to replace. Any wake word solution should match the word locally before creating a connection to AVS and streaming audio data to respond to. Finally, we currently do not support returning text to the user instead of a sound file, although I will forward your interest to the relevant team.
10 |5000

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

Eric@Amazon avatar image
Eric@Amazon answered
From the dev portal: "Push-to-talk devices and applications are currently permitted by the Alexa Voice Service. If you are interested in building far-field voice recognition or using a spoken word to trigger activation of the Alexa Voice Service in your device or application, please contact Amazon ( https://developer.amazon.com/public/support/contact/contact-us)"
10 |5000

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