As already reported by many, the Speech to text feature does not work.

In particular, I am using Korean , and the character recognition works fine in the preview, but all the characters disappear in the process of displaying the final result.

The same bug is being reported so hopefully this issue will be fixed in the next version.