Audio Profanity Remover — Clean Version Maker

Tool for cleaning profanity out of an audio track — a music single, a podcast episode, a voice message or an audiobook. Speech recognition reads the recording, lays every word out on a waveform, and stamps each swear word with its exact start and end so the censor lands on the word and nothing else. The cleaned audio is rebuilt and saved as MP3, WAV, M4A, OGG or FLAC at the bitrate you choose. Recognition covers ten languages — Russian, English, Spanish, Portuguese, German, French, Indonesian, Korean and Arabic; you pick the audio language, independently of the interface language. Processing runs in your browser via WebAssembly.

How to remove profanity from audio

Drop in an audio file — a track, podcast, voice message or audiobook chapter (MP3, WAV, M4A, OGG and more)

Let speech recognition read the recording; progress, time elapsed and time remaining appear as it works

Walk the waveform: click a word in the transcript to spare it, or drag over the wave to silence a missed swear

Set the censor sound — beep pitch and level, silence, a music duck, or your own clip — and listen back in the player

Press "Download clean file" and choose the audio format and bitrate for your release

Bleep the swearing out of a song or podcast straight on the waveform

Audio language

🤬

Drop an audio file here or click to browse

Drag & drop or browse

Loading file...

1×

Click the wave to seek · drag on empty space to mute manually · drag the edge of a red area to move that boundary, or the area itself to shift it · Delete removes the selected one, Ctrl+Z undoes, Space plays and pauses · + and − (or Ctrl and the wheel) zoom in

Keyboard shortcuts

Space: pause / play
J K L: back 3 s · pause · forward (press L again to go faster)
I O: start / end of the selected censor at the playhead; with nothing selected — mark the in and out of a new censor
↑ ↓ · [ ]: previous / next censor
← →: one frame back / forward (Shift — one second)
, .: move the selected censor by 10 ms (Shift — 100 ms)
Shift+K: play the selected censor with some lead-in
Delete: remove the selected censor
Ctrl+Z · Ctrl+Shift+Z: undo / redo
+ − · Alt: zoom the waveform in / out; hold Alt while dragging to turn snapping off

0:00 / 0:00 Found 0 Muted 0 Preview with censor

Transcript

Click any word to mute it or remove the mute.

Will be muted

No profanity found. Click a word in the transcript or drag on the wave to add one.

Censor with

Beep frequency (1000 Hz — TV, 440 — radio) 1000 Hz

Censor volume 50% Auto

The beep sets its own level: 10 dB below the speech around it, as on TV and radio

Repeat if the censor is longer than the sound

Audio file up to 20 MB; the first 30 seconds are used. The sound plays once from the start of each censor — if the censor is longer, the rest is silence.

Word coverage 100%

How much of the word is muted. 100% = the whole word; lower keeps the word edges audible (more recognizable).

What to detect Obscene words (always on) Rude words and insults

Not profanity in the strict sense (insults and crude words), but most people want them censored too — so it is on by default. Untick to censor swear words only.

Euphemisms

Custom word lists (blacklist and whitelist)

Custom words (blacklist)

Never mute (whitelist)

Finished file

Audio is rebuilt with smooth fades; in videos the picture is copied without re-encoding.

Project:

A project is a file on your computer: the transcript, every edit, the settings and word lists. Open it with “Open project” above the upload area — on this computer or another. “With the recording” puts the recording itself in the same file, so you don’t have to pick it separately.

🎬 Cleaning a video instead? Video profanity remover

Features

Made for a music or podcast mix

The recognizer picks out swear words even when they sit under a beat or a music bed, so a song or episode comes back clean.

Edit on the waveform

Every detected word is marked on the audio wave; scrub, click a word in the transcript, or drag across the wave to mute a spot the model missed.

Choose how it sounds

Replace each word with a 200–2000 Hz beep, clean silence, a volume duck under the music, or a sound clip of your own — with smooth fades so there is no click.

Audio export the way you need it

Render the clean version to MP3, WAV, M4A, OGG or FLAC and pick the bitrate — ready to upload to a streaming or podcast platform.

Tune what counts as profanity

Keep obscene roots on, switch rude words or euphemisms on or off, and feed your own blacklist and whitelist for an artist or show name.

Cue sheet of every edit

Export a timecoded list of every muted word as TXT, CSV, SRT or JSON for your records or a label submission.

Why use this audio cleaner

Hunting for swear words by ear in an audio editor means scrubbing the same minute over and over and still missing one before the song or episode goes live. This tool transcribes the recording, pins each word to its timecode on the waveform, and lets you beep, silence or duck it in a click — then hands you a clean MP3, WAV, M4A, OGG or FLAC at the bitrate your distributor wants. Whether it is a radio edit of a single, a brand-safe podcast cut, or an audiobook for a younger audience, the clean version is ready in minutes instead of an afternoon of manual editing.

Frequently Asked Questions

Will it catch swearing in a song with a backing track?

Yes. The models are trained on real recordings and read vocals over music, though a very dense mix can blur a word. Anything it misses you can mute by hand by dragging the spot on the waveform.

Can it find a swear word buried inside a music mix?

That is what it is tuned for. A recogniser aimed at clean speech tends to miss a word sitting under a drum kit and a bass line, which is exactly where the words you need to censor usually are on a single. Everything it finds is marked on the waveform, so you scrub to each one and hear it in place before deciding — and a word it missed can be marked by hand. Nothing is bleeped that you did not confirm.

Does the song stay in sync after cleaning?

Yes. Words are muted in place rather than cut, so the running time is identical and nothing drifts; short fades at each edge keep the audio smooth.

Can I drop in my own bleep sound?

Yes. Switch to "My own sound" and upload a short clip; its first 30 seconds replace each muted word, looping over longer ones, with the level under your control.

Which language does the recogniser understand?

Ten languages: Russian, Ukrainian, English, Spanish, Portuguese, German, French, Indonesian, Korean and Arabic. You pick the audio language in the list above the upload zone (or let it be detected), independently of the interface language — so you can open the site in Arabic and clean an English podcast. Russian runs on GigaAM (trained on 700,000 hours of Russian audio); the other languages run on Whisper. Recognition accuracy varies by language, so review the words that were found — anything missed can be muted by hand on the waveform.

💡 Want us to improve this tool just for you?

We can! Just send us a quick message with your idea. If you'd like to discuss it in detail, leave your email and we'll get back to you. You can stay anonymous.

How do you rate this tool?

1 (1 rating)

Thank you for your rating!

Change my rating

Want to leave a comment?

Want to share more? Leave a comment!

Thank you! Your comment will appear after moderation.

Who is this tool for?

Published 10 Jun 2026 Updated 11 Jun 2026 Author: The Supreme