Create & edit · Free
Remove ums and filler words from video
Find every um, uh and hesitation sound in your talking video and cut them in one click. You see each one in the transcript and keep the ones you like – free, no upload, no watermark.
- No account
- Runs on your device
- Phone & desktop
Remove ums from your video
Stays on your device
Drop a video or audio file
MP4, MOV, MKV, WebM, AVI · MP3, M4A, WAV · phone, WhatsApp and screen recordings, podcasts
Nothing is uploaded. Your file is read, transcribed and cut on this device.
This browser cannot play the file here. Cutting and export still work.
Text
Click a word to jump there. Drag across words (on a phone: tap the first and the last) and press Cut. Ctrl+Z undoes.
Saved you an hour of editing?
Cut Studio is free and has no watermark. A coffee helps keep it that way.
No upload. The speech recognition (OpenAI Whisper) and the cutting run in your browser. Only the speech model is downloaded, once.
Fast on a computer. With a graphics chip (WebGPU) the words of a 10-minute video are ready in one to two minutes. Phones work too, best with shorter clips.
Remove filler words in three steps
Open the video
Cut Studio writes down every word with its time and marks the ums it hears.
Review the marks
Filler words are underlined, voice sounds without a word show as 〰 chips. Click a chip or select a word to keep it.
Export
The preview skips every cut. Export the MP4 – or an EDL/FCPXML cut list for your editing program.
Pauses that feel natural
Cut Studio measures the loudness of your recording every 20 milliseconds and learns the level of the room around you. Only silence longer than your threshold (0.7 seconds to start with) is cut, and a short breath stays on both sides, so the speech keeps its rhythm instead of sounding chopped. Silence before the first and after the last word goes completely.
Finding the ums
Speech recognition is trained to write clean sentences, so it often leaves out an “um” altogether. Cut Studio therefore looks twice: it removes the hesitation words the recognition did write down, and it marks stretches where you can hear a voice but no word was recognized. Those show as small 〰 chips in the text; click one to keep it.
For editors: subtitles and cut lists
Already working in DaVinci Resolve, Premiere Pro or Final Cut Pro? Take the cuts along as an EDL or FCPXML file and import them as a timeline over your original footage, at full quality. The SRT and VTT subtitles are timed to the edited version, so they line up after every cut.
More ways to use Cut Studio
Frequently asked questions
Why does it find ums that are not in the text?
Speech recognition tends to tidy up sentences and leaves many ums out. Cut Studio also listens for stretches with a voice but no recognized word; those are usually hesitations, a laugh or a cough, and you decide for each one.
Which filler words does it know?
Um, uh, erm and hmm in English, äh and ähm in German, euh and hum in French, eh and ehm in Spanish, plus the sound check that works in every language. Real words such as “like” are never cut automatically.
Is Cut Studio really free?
Yes. No account, no watermark, no limit on how long your videos are or how many you edit. Fair ads below the tool and voluntary coffees pay for the site.
Are my videos uploaded?
No. Your file is read, transcribed and cut in your browser. Our server only delivers the page and the speech model; it never sees your recording.
Which languages does it understand?
Whisper recognizes about 99 languages, from English, German, French and Spanish to Japanese and Arabic. Cut Studio detects the language by itself, or you choose it.
Does the export lose quality?
The edited video is encoded again as H.264 at a bit rate close to the original’s. If you want to keep your original footage untouched, export the EDL or FCPXML cut list and render in your editing program.
How long can a video be?
On a computer, hour-long recordings work. Phones have less memory, so clips of up to about 15 to 20 minutes are the safe choice there.
Take a break
