Filler Word Remover

Remove Filler Words from Video and Audio

QUICK ANSWER

Upload a video or an audio file and every um, uh, er and hmm is found by AI speech recognition and marked for removal in the transcript. Keep any cut you want, cut extra words yourself, hear the result, then export an MP4, MP3 or WAV. Free, no signup, up to 500 MB and 30 minutes.

  • Finds um, uh, er and hmm
  • Review every cut in the transcript
  • Cut any other word yourself
  • Tighten pauses too
  • Video or audio in, MP4, MP3 or WAV out
  • No signup
  • Up to 500 MB

Processed on our server. Your upload and the edited file are deleted within 24 hours.

See it in action0:49

This walkthrough is itself a Clipy recording. Watch it on Clipy with its transcript and summary

Why this tool exists

  • Your file uploads straight to our storage through a private, expiring link.
  • Speech recognition and the cut render run on our server, not in your browser.
  • The upload and the edited file are deleted within 24 hours.
Trusted by creators at startups, agencies, and Fortune 500 teams.
This tool is free to use, with no signup.

How it works

  1. 1

    Upload a video or audio file

    MP4, MOV, WebM, MP3, M4A or WAV, up to 500 MB and 30 minutes long.

  2. 2

    Check the crossed-out words

    The transcript comes back with every filler word crossed out. Select a crossed-out word and choose Restore words to keep it, select any other words to cut them, and play the edit with the cuts skipped.

  3. 3

    Export

    Choose MP4 to keep the picture, or MP3 or WAV for audio, and download the result.

Which filler words are removed

Hesitation sounds: um, uh, er, erm, ah, hmm and mm, including drawn-out spellings like umm and uhh. Words such as "like", "so", "actually" and "you know" are never cut automatically. They are filler in some sentences and meaning in others ("I like it"), so cutting all of them would break sentences. Select the ones that are filler in the transcript and cut them yourself.

How a filler word is cut

Each filler word is cut from its first sound to its last, plus at most 80 milliseconds of the pause on each side, and never more than half of that pause, so the words around it keep their natural spacing. The picture and the sound are cut at the same instants, so a talking head stays in sync. Long pauses are left alone unless you turn on Pauses to remove; then gaps longer than 0.4, 0.7 or 1.2 seconds are shortened to 0.15 seconds of air on each side.

Why you review the cuts first

Speech recognition is good, not perfect. It can hear "uh" where someone said "a", or miss a quiet hesitation. So nothing is cut until you have seen it: crossed-out words are cuts, a red strip under the player shows where each one falls, and clicking a red block (or selecting a word and choosing Restore words) brings it back. Turn on Skip cuts while playing to hear the edit before anything is rendered. If the export is not right, change a cut and export again from the same upload.

For podcasts, interviews and voiceovers

Audio files work the same way as video. Upload an MP3, M4A or WAV episode, review the transcript, and export MP3 or WAV. For a video podcast, export the MP4 for YouTube and an MP3 of the same edit for your feed.

Filler word removal in other tools

Checked in October 2026 on each vendor's own pages.

ToolFiller word removal on the free planAccount needed
ClipyYes, with a transcript review, no watermarkNo
DescriptLimited; free exports carry a watermarkYes
VEEDNo, Magic Cut is a premium featureYes
RiversideNo, a Pro featureYes
LoomNo, only on the Business + AI planYes

Related tools

To cut the dead air without touching the words, use the silence remover. Get a plain transcript with video to text, captions with the subtitle generator, and even out levels with normalize video audio. Recording your next walkthrough? Clipy's screen recorder gives you a share link and a transcript as soon as you stop.

Common questions

How do I get rid of filler words in a video?

Upload it here, check the crossed-out ums and uhs in the transcript, and export. Every filler word the speech model hears is marked for you; bring back any you want to keep before exporting.

Can it remove "like" and "you know"?

Not automatically, because those words often carry meaning. Select them in the transcript and they are cut like any other word.

Does it work on audio files and podcasts?

Yes. Upload an MP3, M4A or WAV file and export MP3 or WAV. From a video you can export an MP4 or just the audio as MP3.

Which languages does it work with?

The transcript works in around 40 languages, and you can cut any word in any of them. Automatic filler detection looks for English-style hesitations (um, uh, er, hmm), which many languages share; anything else, select and cut yourself.

Is it free?

Yes, with no signup. Anonymous visitors get six runs an hour and a free Clipy account gets twenty. Up to five more exports of the same upload within two hours, after changing a cut, do not count as more runs.

What happens to my file?

It is uploaded to our storage, transcribed and cut on our server, and deleted along with the edited file within 24 hours.

Tool not working the way you expect?

Tell us what broke. We resolve every report within 24 hours.

Send feedback →

More free tools

See all free tools.

Doing this for an AI workflow?

Clipy recordings are agent-readable out of the box. One link carries the video, keyframes, and transcript.

Clipy for AI agents