I maintain a web chat product and would love to replace the front-end with chatkit but our current product has a voice mode. When voice mode is enabled we read the STT and update the chat window as a record of what is being said on the voice call.
I don't see anyway to make that work with the current api. It would be great if you exposed some methods to let us control adding user and assistant messages without triggering any requests.