So, a common screencast approach is to record a screencast, capturing voice/screen with multiple tries and allows. The edit that screencast to get the screen the way you want it, and then re-record the audio while playing the screen and the edited script which matches the timing of the screen cast. This is important as you may be highlighting lines of code or using the mouse to focus on items. Timing must match the screen.

So, it would be handy to have a teleprompter that would play the script with that timing - and allow recording of face/audio.

It should allow the text to be near the camera, and the playing video below, so the speaker knows what the viewers are seeing.