How to remove filler words in CapCut, Descript, Premiere Pro, DaVinci Resolve and Hilite

Modern filler word removal works the same way everywhere. The software transcribes your recording, marks words like um, uh and er in the transcript, and lets you delete them. Because each word is linked to its moment in the audio or video, deleting the word cuts the media underneath.

That means the quality of the result depends on two things: how accurate the transcript is, and how carefully you review the cuts. The steps below cover the main tools. Menus change between versions, so if a button has moved, look for the transcript or text-based editing panel.

In recent versions of CapCut desktop, transcript-based editing can detect filler words. Generate a transcript for your clip, open the transcript editing view, and look for the option to detect or remove filler words. CapCut highlights the ums and uhs, and you can delete them individually or together.

Feature availability, and whether it needs a paid plan, varies by version and region. If you do not see the option, update CapCut or use the transcript to search for um and uh manually. If you are working on audio first, see how to add an audio file to CapCut.

Descript also edits through the transcript. Once your recording is transcribed, use its remove filler words option, which detects common fillers and lets you choose which types to remove. You can review each one in the script before applying the change.

Check the result in places where a filler sits next to a breath or between two words spoken quickly. Those are where cuts most often sound clipped.

Premiere Pro's text-based editing can transcribe a sequence and detect filler words in the Text panel, where you can select and delete them, along with long pauses. Because it cuts the video too, review jump cuts on camera and cover them with B-roll or a zoom if needed.

DaVinci Resolve's recent versions include text-based editing with transcription and silence removal. If filler detection is not available in your version, search the transcript for um and uh and delete those words, which removes the matching section of the clip.

Hilite is built for spoken audio. Record or upload a file, and the transcript appears with filler words detected. Remove all of them in one pass, or review and keep the ones that sound natural. Edits are non-destructive, so you can undo any cut. Filler word detection currently works in English.

For podcasts and voice recordings, this is usually the fastest route. If you are making a video, clean the audio in Hilite, then sync it in your video editor.

Removing every filler can make speech sound rushed and slightly robotic, because natural speech has small hesitations. Listen back to sections with many cuts and restore any that leave a sentence feeling choppy.

Watch for fillers that are part of meaning, like "like" in "it looks like rain", or that carry a breath. Automatic tools sometimes catch real words. A quick review after the bulk removal catches these in a few minutes.

Editing cleans the recording but not the habit. To use fewer fillers when speaking, replace them with a short pause, prepare the first sentence of each section, and slow down slightly. See how to stop saying um and how to stop saying like.