How to bleep out words in a video
You have a finished video. Somewhere in it there are a few swears, and they need to be gone before it goes up. There are two ways to bleep out words in a video: do it by hand in whatever editor you already have open, or let a transcript find the words and drop the bleeps for you. Which one is right depends on how many swears there are and whether you know where they are.
The other half, which most guides skip, is that a bleep is easy to place and easy to place badly. Where it starts, where it ends and how loud it is decide whether the word is actually hidden. Both halves below.

The short version
One or two swears in a short clip, and you know where they are: do it by hand. It is free, it takes a few minutes, and you keep control of the timeline.
More than a handful, or a long recording where you would have to scrub the whole thing to find them: use a transcript-based tool. The bleeping is not the cost. Finding word 14 of 23 in a 40 minute video is the cost. (Want to compare tools first? We did that in best apps to bleep out swear words in a video.)
The automatic way: transcribe, pick the words, export
Every automatic video bleeper works the same way underneath. Speech recognition turns the audio into a transcript with a timestamp on every word, you mark the words to censor, and the tool mutes each one and lays a sound over the gap.
In Instableep that is: upload the video (MP4, MOV, WEBM, MKV, or a plain audio file), wait for the transcript, apply the default curse-word preset or type your own words, choose a bleep or silence, listen back, export. It runs in the browser, phones included, and the library keeps the file and transcript so a re-edit next week does not mean transcribing again. It is built for speech, not songs.
The step people skip is the listen-back. Speech recognition is very good and not perfect. Quiet mumbles, heavy accents, two people talking at once and the odd innocent word that sounds like a swear all trip it. Read the transcript before you export instead of trusting the preset blindly. More on that in AI profanity filter for video.

The manual way in any editor
The four steps are the same everywhere. Find the word. Split the clip just before and just after it. Mute that slice (mute, not delete, or everything after it slides out of sync). Drop a bleep sound on a track under the gap and trim it to fit. Export.
Only the menu names differ:
-
CapCut: split twice, set that piece’s Volume to 0, then Audio, Sound effects, search “beep” and drag one under the gap. Walkthrough in how to bleep out cuss words in CapCut.
-
Premiere Pro: transcribe the sequence, filter the Text panel for the swears, and delete the matches with Lift ticked so nothing shifts. Then lay one long bleep tone on a second track and punch the dialogue over it, so the tone only survives in the holes. Details in how to bleep out words in Premiere Pro.
-
DaVinci Resolve: blade around the word, pull the clip gain all the way down, add a track and drop a bleep file on it. Resolve can also generate the tone in Fairlight, covered in how to bleep out words in DaVinci Resolve.
-
iMovie: Modify, Detach Audio, then Modify, Split Clip at each end of the word, mute that piece with the Volume button, and drag a beep in from Audio & Video. iMovie ships no censor beep, so you bring the file.
Pixabay’s censor beeps are free to use in your own video, no credit needed; you just cannot resell the beep on its own. Or generate a one second beep in Audacity. Both in bleep sound effect for swearing.
Where the bleep starts and stops
This separates a clean bleep from one that fools nobody, whether a person or a tool places it.
-
Start early. A bleep aligned to the visible bump in the waveform starts on the vowel. The “f” or “sh” in front is quiet, sits before the bump, and gets through. Your viewer hears “f-BEEP” and fills in the rest.
-
End late. A trailing “-ck” or “-it” is just as readable. Overshoot on both sides by a fraction of a second, on purpose.
-
Fade both edges. A pure tone spliced hard into speech clicks at the cut, and the click is more distracting than the swear was.
-
Keep it under the dialogue. Broadcast beeps are loud because television wanted them loud. On headphones a bleep that peaks the meter makes people flinch. Pull it down to roughly the level of the speech around it.
In an editor that is four adjustments per word. In a transcript tool it is three settings (volume, fade, margin) applied to every bleep at once.
Bleep or silence
The classic beep is the broadcast convention, so it reads as “TV censorship” and suits content where the censoring is part of the joke: comedy, gaming, reactions. For an interview or anything with a serious register, a muted gap is less intrusive, and live broadcasters mostly mute rather than beep for the same reason. Novelty sounds (a quack, an air horn, a record scratch) are a style choice: pick one and use it every time, or it reads as an accident. Longer argument in bleep, silence, or quack.
Does bleeping hurt the video quality?
It can. A timeline editor rebuilds the whole video to change one second of audio, and every rebuild costs a little picture quality. The fix is to leave the video stream alone: Instableep’s paid exports copy the video untouched and only rebuild the audio, so the picture comes out exactly as it went in (WEBM comes back as MP4). By hand, export once from the master, never from an already exported copy.
What a bleep will not fix
If the speaker’s mouth is on camera in close-up, the word can still be lip-read. Covering the mouth is a video-editor job, and no audio tool does it.
On monetization, relax a little. YouTube counts bleeping and muting as obscured profanity, which sits in the tier that can earn full ad revenue. Titles and thumbnails are the real risk, and the old “first seven seconds” rule was relaxed in July 2025. Details in YouTube profanity rules.
For the no-editor route, see the online swear word remover.
Try Instableep for free, 20 minutes of transcription a month, no card, watermark on the free export. A $4.99 pack of 60 minutes removes the watermark for good.