Dear Suzannah

Edit the transcript, not the waveform.

Descript turns your recording into a transcript. You delete words, sentences, or sections in the text and the audio edits automatically. No more guessing where to cut on a timeline.

The real use case

You recorded an interview or voiceover and need to remove filler words, long pauses, and an off-topic section in the middle. Instead of editing audio waveforms by hand and risking bad cuts, you import the recording into Descript, let it generate a transcript, and edit the text directly. Deleting a sentence in the transcript removes the matching audio. You clean up the content, export the result, and get a polished recording without learning traditional editing software.

No prompt needed: use this workflow instead

Descript is not a blank prompt tool for this use case. Use the transcript-based editor instead. Follow these steps after importing your recording.

1. Open Descript and create a new project, then drag your audio or video file into the project window to start the transcription.
2. Wait for the transcript to generate, then read through it to find the filler words, long pauses, and sections you want to remove.
3. Highlight any word, sentence, or paragraph in the transcript that you want to cut and press Delete — the matching audio is removed automatically.
4. To remove filler words like um and uh in bulk, use the Remove Filler Words option from the Edit menu and review what it flagged before accepting.
5. Play through the cleaned-up recording by clicking the Play button in the transcript to check that the edits sound natural and there are no abrupt gaps.
6. If a cut creates an awkward pause, add a short crossfade by selecting the join point and applying a fade transition from the toolbar.
7. When the recording sounds clean, go to File, then Export, choose your audio format (WAV or MP3), and save the finished file.

Make the result less generic

• Read the full transcript first and mark the sections to cut before deleting anything, so you do not lose context.
• Review the filler-word removal list before accepting it — some flagged words may be intentional for tone.
• Play through after each major cut to catch awkward gaps that need a crossfade.
• Export at the bitrate your publishing platform requires instead of using the default.
• Keep a copy of the original file in case you need to restore a section you deleted.

Quick human check

• Does the recording sound natural after the cuts, with no abrupt gaps or clipped words?
• Did you remove all the filler words and pauses you identified in the transcript?
• Is the off-topic section completely gone, or did a fragment remain?
• Does the exported file play correctly at the right format and quality?
• Did you keep the original recording in case you need to undo an edit?

Sources and further reading

• Descript — official site
• Descript transcription
• Descript pricing