Claude wrote my podcast script. Descript cut the ums out of me reading it. Neither did the other's job

My solo podcast had a production problem: writing a decent script took one evening, and cleaning up my spoken delivery took another. Two tools, each owning one evening. Claude owns the writing now, Descript owns the aftermath, and the whole thing fits in one session.

Claude's script work is the better half of my show, honestly. I feed it the topic, my opinion in bullet form, and the shape of past episodes I liked. It drafts with actual structure — a cold open that isn't 'so today we're going to talk about,' transitions that connect instead of just following. Then I argue with it, which is where my voice gets in.

Descript handles what happens after the mic. Read the script, drop the WAV in, and it transcribes everything so I edit audio by editing text. Delete a sentence in the transcript, it's gone from the audio. The filler-word removal is one click and it kills every 'um' without making me sound sped-up. My first reaction was suspicion. My listeners' reaction was asking if I'd gotten a producer.

The handoff detail that matters: Claude writes for reading, not speaking. Left alone, it produces beautiful sentences that turn into tongue-twisters out loud. My fix is a final prompt — 'mark anything over twenty words long' — and reading the draft aloud once before recording. Descript will clean the audio, but it can't un-awkward a sentence.

They don't overlap and they don't compete: Claude is the writer, Descript is the editor, and I'm just the mouth. If you host a talking show and still scrub waveforms by hand, Descript alone will feel like a raise. Add Claude and you get your evenings back too.

Frequently asked questions:

Q: Can Claude edit my podcast?

A: No - it works with text only. Draft and plan with Claude, then edit the recording in Descript.

Q: Can Descript write my episode script?

A: Only basic AI drafting. For research-heavy scripts and show notes, Claude is the stronger writing partner; Descript is where the script becomes audio.