WizardCut TUI

Edit videos by deleting text. Whisper transcribes your video, you delete what you don't want, FFmpeg cuts it out.

Installation

git clone https://github.com/oidz1234/wizard-cut-tui.git
cd wizard-cut-tui
pip install -r requirements.txt

Requirements

• Python 3
• FFmpeg
• mpv (for live preview)
• vim or neovim

Usage

# Basic - transcribe, edit, cut
python wiz.py -f video.mp4

# Use a faster/smaller model
python wiz.py -f video.mp4 -m tiny

# Specify output path
python wiz.py -f video.mp4 -o output.mp4

# Pin a language, or omit this flag to let Whisper auto-detect
python wiz.py -f video.mp4 -l en

# Show shorter silences as editable markers
python wiz.py -f video.mp4 --silence-threshold 0.6

# Disable live preview
python wiz.py -f video.mp4 --no-preview

How it works

1. Whisper transcribes your video with word-level timestamps
2. The transcript opens in vim/nvim
3. Delete any words, sentences, or silence markers you want to remove
4. Save and quit - the video is processed with the cuts applied

Live Preview

Preview is enabled by default when mpv and vim/nvim are available. An mpv window opens alongside your editor:

• Move your cursor through the transcript - mpv seeks to that timestamp
• Press F5 to play/pause - the cursor follows playback in real-time
• Save (:w) to update the preview - deleted segments are skipped during playback
• Press ? for controls help

Insert mode is disabled to prevent accidental text addition.

Notes

• For best results, delete whole words or sentences rather than single characters
• The default Whisper model is medium - use tiny for speed or large for accuracy
• If mpv cannot start, WizardCut falls back to the plain editor flow
• Edited video is saved to your current directory as {name}_edited.{ext}