Language identification

Is there a way to use the Riva framework to do language identification directly from audio signal, or are there plans to include such feature in the near future? Thanks!