stt.md 1.4 KB

Speech To Text (STT)

This chapter describes how STT are working.

The STT are programs or API that converts the voice to text. Each of them have a specific configuration and support multiple voice and/or languages.

The configuration for each STT must apply in the settings.yml file.

The syntax used is YAML.

General defaults

The setting.yml defines the STT you want to use by default

default_speech_to_text: "type default STT here"

Then, still in the settings.yml file, each STT must set up its configuration following 'speech_to_text' tag :

speech_to_text:
   - STT1:
      STT1parameter1: "value option1"
      STT1parameter2: "value option2"
   - STT2:
      STT2parameter1: "value option1"

Current Available STT

apiai bing google houndify witai

Full Example

In the settings.yml file :

default_speech_to_text: "google"
speech_to_text:
  - google:
      language: "fr-FR"
  - wit:
      key: "B5JI3YUSLYOYWNIZRNBVM34XUODME2K"
  - bing:
      key: "9e48dert65904838bedc11aea6b36fb0"
  - apiai:
      key: "e0cbff145af44944a6b9f82c0668b527"
      language: "fr"
  - houndify:
      key: "6ej90T7qAV74OYXk4X4vI2Xhk7wPsJu4aEZ0G5Ll-BMmV1JGtFpCxtSH9SmTY4G3bpEJ7a5y_GTQid-CAKI6vw=="
      client_id: "lM2JXeaSticbSo9-llczbA=="