I have been trying to use Google Speech-to-text v2 in node.js to recognize streaming input. For those who have worked on the same, they must know that this version requires us to create a ‘recognizer’ which depends on the location and model that we want to transcribe the language with. For a lot of languages, we have the global location. For some languages, global is not supported. Some examples are: bn-IN and yue-Hant-HK.