This tool removes filler words and silence from video or audio. Upload a talking-head take, podcast, lesson, or voiceover. Quiet stretches are measured in the file. Words like “um,” “uh,” “you know,” and “I mean” are taken from the transcript. Those ranges are cut and the remaining speech is joined in the original order. You download a shorter file and a list of what was removed.
Useful for Creators, teachers, anyone who records themselves
Or use the general upload for a transcript, summary, chapters, and chat. All tools.
Run this tool
Upload one recording. We cut silence and filler words and return a shorter file.
A talking-head video is rarely the length people think it is. You sat down for twenty minutes. Several of those minutes are waiting, throat-clearing, and filler. This tool does two different jobs on the same file. First it looks for quiet stretches in the audio itself — dead air and long pauses where almost nothing is heard. Then it looks at the transcript for filler such as “um,” “uh,” “er,” “erm,” “hmm,” “ah,” “huh,” “you know,” and “I mean.” Those times are merged and removed. Useful sentences stay in the order they were said. We do not shuffle topics, rewrite lines, or build a highlight reel from a title. If overlapping speech sits on top of a quiet stretch, that overlap can be cut with the gap. Soft background music can look like silence, or it can hide a pause so the gap is not measured. A noisy room can make silence detection fail; filler words from the transcript are still listed.
Creators, teachers, podcasters, and anyone who records themselves and wants a tighter file before they publish or send it.
You recorded a first take and hear thinking sounds between every sentence. Upload that video or audio. The transcript marks those filler cues; those times are cut with any long quiet around them. You download a shorter file and a list of each filler line. The sentences stay in order, so the take still sounds like you — just without the stalls.
You stopped to check notes or wait for a slide. Those gaps make the video feel slow even when the words are fine. Upload the file. Quiet stretches in the audio are measured and removed. You get a tighter cut and rows labeled as quiet stretches with how many seconds left. Short natural gaps can remain; this is not a cut of every breath.
Two people leave space between answers. You do not want to rewrite the conversation. Upload the episode. Silence between turns is cut; filler in either speaker’s lines is cut if it appears in the transcript. You leave with a shorter episode and a cut list. Overlapping laughter or crosstalk that sits inside a “quiet” range can disappear with that range.
A 20-minute screen-and-face recording has four minutes of waiting. Upload it as video. After the cuts you download a shorter MP4 (or MP3 for audio-only) you can post or hand to an editor. Music under the talk can hide pauses or be treated as silence — listen to the result before you publish.
A teacher uploads a 20-minute lesson. The audio has three long waits while a slide loads. The transcript also has several “um” and “you know” cues. Those ranges are removed and the remaining speech is joined. The cleaned file might land around 16 minutes. The cut list is illustrative of the shape you see — not a recording we processed for someone else.
Quiet stretch 03:12–03:48 (no speech) 36.0s removed Quiet stretch 11:02–11:21 (no speech) 19.0s removed Filler 04:05 “So, um, the next part is the formula.” Filler 09:40 “You know, that’s the bit students miss.” Cleaned file: lesson-cleaned.mp4 Speech kept in the original order.
A shorter audio or video file with speech kept in order, plus a cut list of quiet stretches (with seconds removed) and filler lines (with the spoken text and time). Metrics show how many stretches were removed and about how many seconds left. If nothing qualifies, we say so and do not invent a cut.
MP4, MOV, or WEBM video, or MP3, M4A, or WAV audio. Upload from your device. We do not fetch files from other sites.
Silence measurement can fail on a noisy file; filler from the transcript still runs. Overlapping speech inside a quiet range can be cut. Background music can be mistaken for silence or can hide dead air. We do not separate speakers, and we do not cut coughs, repeats, or false starts unless they match the filler list. Short natural pauses can remain.
Yes. Upload the file. Quiet stretches and transcript filler are cut, then you download a shorter video or audio file. You do not open a timeline.
The transcript is checked for forms such as um, uh, er, erm, hmm, ah, huh, uh-huh, “you know,” and “I mean.” Other habits — repeats, false starts, “like” — are not on that list unless they sit inside a quiet stretch.
No. Short natural gaps can remain. We target measured quiet stretches and filler cues, not every breath.
Overlapping speech is not split by speaker. If it sits inside a range marked quiet or filler, it can be removed with that range.
Sometimes. Soft music can look like silence to the loudness check, or it can hide a pause so the gap is not cut. Listen to the result.
Quiet rows show “(no speech)” and how many seconds were removed. Filler rows show the transcript line and its time.
No. This keeps the talk and removes waste. To keep only a target length of the strongest minutes, use the shorten tool.
Guide: How to cut dead air and filler words without opening an editor
Or use the general upload for a transcript, summary, chapters, and chat. All tools.