Edit a podcast episode in Descript by editing text — remove filler words, silences, and mistakes without touching a timeline.
Click 'New Project', drag in your audio or video file, and wait for Descript to transcribe it. Transcription takes roughly 1 minute per 10 minutes of audio.
Read through the auto-transcription and fix any errors. Click a word to hear the corresponding audio. Accuracy is usually 90–95% for clear audio.
Go to Actions → 'Remove Filler Words'. Descript highlights every 'um', 'uh', 'like', and 'you know'. Preview before deleting — some fillers aid natural flow.
Use 'Remove Silence' to automatically cut pauses over 0.8 seconds. This alone cuts most podcast episodes by 10–15% with no manual work.
Find a mistake in the transcript, select the words, and press Delete. The audio is cut at that point. No need to scrub a waveform.
Descript can clone your voice (requires 10 minutes of training audio). Type a corrected word and Overdub generates it in your voice, seamlessly replacing the original.
Export as MP3 at 128kbps for voice-only podcasts. Enable 'Studio Sound' (AI noise removal) before exporting to clean up background noise.
Free (1 hour transcription/month)