A bug was resolved where audio responses were showing wrong and very transcripts.

Audio response transcripts now work for over 90 languages once again.

The audio transcriptions handle multiple languages within the same response as well.