Two or three approaches here:
#1 Listen for livestreams eg., Facebook, Youtube, LinkedIn, and automatically in-take the feed recording best quality video and transcribing the audio.
Use cases would include a fast turn around on editing into a shorter video, a podcast, Etc.
Also collaboratively start fixing up the edits, etc.
#2 Allow a project / drive to publish an rtmps URL/ API end point so that someone could live stream into Descript.
#3 This would be an expensive Pro feature but in-take the RTMPS stream directly from the sender (eg Zoom, OBS), then resend to various other steams (Youtube, Facebook, Etc.)
Use case: When an app like Zoom only allows 1 stream,. or using an app like OBS naively would use too much bandwidth or CPU (Or both).
Use Case near term: To allow a live event to appear on multiple streaming services from a single feed (while in-taking the stream for transcription)
Use Case long term: Allow real-time AI transcription +/- near live human edits.
Add in a delay (ie 60 seconds) and send a transcript along with the stream. (ie CEA-608 standard closed captions to Facebook via the Live API).