Using ios dictation will trigger beforeinput and input events on DOM page. However, the words will be redundant.

For example, when I speak 'Test Voice Input', I may got continuous input events with data as following:

1. 'Test'
2. 'Test Voice'
3. 'Test Voice Input'

I just want the 3rd result but the above 3 events have no individual property to distinguish them from each other.

Does ios offer some other events to listen to the dictation status?

(I have tried compositionstart and it wasn't triggered)