How to remove filler words from a video

Um, uh, like, you know. They creep into every take and they make a good clip sound unsure. Here is how to cut them out on your phone, cleanly, without shipping your footage to a server.

Updated July 2026 · by the Reelvana team

Download on the App Store Get it on Google Play

Filler words vs silence: two different cuts

People often lump these together, but they are not the same job. Silence removal deletes the quiet gaps where nobody is talking, the dead air between sentences. Filler removal deletes spoken sounds that are technically words but carry no meaning: um, uh, er, like, you know, so. A silence cutter will not touch an um, because an um is not silence. To remove filler you need to work with the words themselves. If you want the silence side of this, see the guide on removing silence from a video; here we focus on the spoken filler.

The transcript approach

The clean way to remove filler is to edit by transcript. Reelvana transcribes your clip on the device by default, using on-device speech recognition, so you get the words with their timings. When you delete a word from that transcript, Reelvana cuts the exact slice of video where it was spoken. Deleting "um" removes the um and nothing else. Because you decide what goes, you never lose a real word to an over-eager filter, which is the usual failure of automatic filler removers.

The steps

1. Bring your clip in. Record in Reelvana or import an existing video.
2. Transcribe on device. Generate the transcript so the app has your words and their timing. Your video never leaves your phone. It runs on device by default, and an optional Cloud model can send just the audio track for higher accuracy, never your video.
3. Delete the filler. Scan the transcript for um, uh, like, and the rest, and delete them. Each deletion cuts the matching video with it.
4. Run silence cleanup. Filler often sits inside a little pause. Run the on-device cleanup with a Podcast or Tight preset to close those gaps, with lead and tail padding so the cuts breathe.
5. Preview and export. Check the timeline, then export. Every edit composes into one decision list and renders once, so the result is seamless.

Why do it on device

A lot of filler-word tools upload your video, run the analysis on a server, and send it back. Reelvana never uploads your video. The transcription, the cutting, and the cleanup run on the phone by default, there is no ad SDK, and it works in airplane mode. If you want the best accuracy, especially in Arabic, an optional Cloud model can send just the audio track for transcription, never your video. The same transcript editing works in English and Arabic, with Arabic laid out right to left.

One part of a bigger cleanup

Removing filler is usually one step in tightening a clip. In Reelvana it lives next to the teleprompter, silence cleanup, animated captions, timeline editing, background blur, watermarks, and sharing, so you can record, cut the ums, caption, and post without leaving the app or exporting between steps.

Common questions

How do I remove um and uh from a video?

Transcribe the clip on device, delete the filler word in the transcript, and the matching video is cut with it. Then run silence cleanup to tidy the pause.

Is this the same as removing silence?

No. Silence removal cuts quiet gaps; filler removal cuts spoken um and uh. They pair well: cut the filler, then close the gap.

Does it find filler automatically?

You get an on-device transcript and cut any word by deleting its text, so it is fast but precise. You choose what to cut, so real words are safe.

Will the cut sound clean?

Yes. Edits render as one pass, and cleanup adds padding so cuts breathe. You can preview on the timeline first.

Is my video uploaded?

No, your video is never uploaded. Transcription, editing, and cleanup run on device by default, so it works offline. If you turn on the optional Cloud model for higher accuracy, it sends just the audio track, never your video.