v0.4.0-alpha · open · local-first · live

Clean My Video.

Drag in any clip. Drop out a clean copy. CMVideo finds every swear and slur with on-device speech recognition, then silences, beeps, or replaces them with a Microsoft Sam style TTS - whichever you pick.

Download choose your OS below View on GitHub

100% local. No accounts, no telemetry, no clip uploads. Audio never leaves your machine.

●

Local & private

Whisper-class speech-to-text runs on your machine. Audio is never uploaded; the only network calls are when you ask CMVideo to download a video from a URL.

▲

Drag-drop or URL

Drop one or many files, or paste a YouTube / yt-dlp-supported URL. Pick the output format (MP4, MOV, MP3, WAV, OGG) and quality. Batches run unattended.

◆

Silence, Beep, or Fun

Replace flagged words with crisp silence, a classic 1 kHz censor beep, or a Microsoft Sam style TTS overlay that turns a 7-hour stream into something genuinely watchable.

Get CMVideo 0.4.0-alpha

Source release. Requires Python 3.10+. The bundled install.sh / install.ps1 sets up ffmpeg and a private venv with the heavy ML deps on first launch - usually 1-3 minutes.

L
Linux
.tar.gz · install with ./install.sh
W
Windows
.zip · right-click install.ps1 → Run with PowerShell
M
macOS
.tar.gz · works via the Linux instructions untested

Checksums and full release notes: github.com → releases → v0.4.0-alpha

Three steps

  1. 1

    Drop a file or paste a URL

    MP4, MOV, MP3, WAV, OGG locally - or any YouTube / 1,800+ site URL supported by yt-dlp.

  2. 2

    Pick what to remove

    Swears, slurs, or both. Choose Silence, Beep, or Fun TTS. Optionally save a full transcript.

  3. 3

    Hit Censor

    CMVideo transcribes, finds every match (including phonetic / leet-speak variants), and writes the cleaned file next to the original.