The Descript Clip Factory: One Podcast, Twenty Social Clips, One Hour
Podcast clips used to be an editor's afternoon. This Descript workflow turns each episode into twenty subtitle-burned verticals before lunch.
Every podcast episode contains twenty social clips and almost nobody cuts them, because clipping feels like editing work. The Descript workflow I use turns it into a find-and-refine job: one hour per episode, twenty verticals, burned-in subtitles, ready to schedule.
Step one is the harvest, and it happens in the transcript, not the timeline. Play the episode at 1.5x while reading along and mark every line that could stand alone: a strong claim, a number, a disagreement, a punchline. Twenty marks on a 90-minute episode is realistic. You are mining sentences, not editing footage, which is why it is fast.
Step two is the cut discipline: each clip is 20 to 45 seconds, starts on the interesting words (cut the 'so what happened was'), and ends on a landing rather than a trail-off. Descript's transcript editing makes each cut a text deletion, and the filler-word removal cleans the rest in one click.
Step three is the vertical conversion, which is where Descript's layouts do the work: apply the vertical layout, position the face crop once per speaker, let captions burn in automatically, then fix the two or three caption lines the auto-sync mangled. Consistency across the twenty clips matters more than polish on any one.
The distribution math is the real argument: twenty clips at an hour of work means three minutes per asset, and even one breakaway clip per episode pays for the hour. My system runs on publish day, the clips drip across two weeks, and the episode page gets a 'clips' section that quietly doubled its average time on page.