Yesterday, the Descript released a new Dynamic Text feature which can display the active Speaker, Marker, Transcript word, Composition Name, or running Timer.
The interface is fluid, and there's no scripting required to get the job done.
I've been waiting for this feature (and waveforms) for a long, long time, so I'm excited to build a Getting Simple template and release clips from previous and future episodes on YouTube and social media.
It'd be great to be able to link the visibility (or opacity) of other canvas elements, such as images or shapes, to show only when a given speaker is talking.