Family member uses Kagi Translate extensively for Japanese to English and vice versa. They often have bizarre audio where the words are repeated multiple times, punctuation is said outloud, and/or other weird stuff like the attached audio file.
They report that it seems to happen most often when text is pasted and for the English speech. Note, couldn't attach original .wav file so converted to mp3. Image of what caused it attached, too. I have asked them to send more examples and I will upload this post when I receive them.
They are using Kagi translate as PWA on iPhone
Get accurate pronunciation audio
