you can now use the OpenAICallTranscription node to convert speech to text setting temperature and expected language. Extended info for recognized text is supported. There is no temporary files to be created as USoundWave pointer is using as a parameter.