Created on 07 September 2021
icon_unfollowing
Login to follow
google-speech-to-text

Google Speech_To_Text

Version 1.0.2 (Compatible with OutSystems 11)
Uploaded on 08 September 2021 by 
google-speech-to-text

Google Speech_To_Text

Documentation
1.0.2

Prerequisites: An active Google Cloud Account, see https://cloud.google.com/gcp/getting-started for more information.


Simple speech recognition

In order to simply run speech recognition, only email and private key need to be populated with the Google service Account JSON information.

So, basically the steps are:

  • Create Google Cloud project (optional)
  • Inside your Google Cloud project:
    • Activate Text-To-Speech API
    • Create Service Account (APIs and Services -> Credentials)
    • Add key to Service Account, download JSON file, save in a secure place (you can only download this once)
  • Open the JSON file, copy content of private_key and client_email into the Site Properties of the Connector Module

Speech model adaptation

  • Inside your Google Cloud project
    • In "IAM & Admin" -> "Manage Resources" , select your project and "Add Member".
      • Find your Service Account and assign the Role "Cloud Speech Client" 
    • Navigate to "Speech-To-Text" -> "Model Adaptations"
    • "New Resource"
    • Enter the phrase set name and the phrases you would like to "boost"
    • Copy the phrase set name into the Site Property, separating several entries with comma


Notes:

phrase sets are added to the API request as PhraseSetReferences, see https://cloud.google.com/speech-to-text/docs/reference/rest/v1p1beta1/RecognitionConfig#speechadaptation for more information.

This component uses the Google API Token extension to fetch the API token. Currently, there is no caching mechanism, the module fetches a new Token for each request.


Support options
This asset is not supported by OutSystems. You may use the discussion forums to leave suggestions or obtain best-effort support from the community, including from  who created this asset.
Dependencies
Google Speech_To_Text has no dependencies.