WizardCut: Text Based Video Editor

Select text to remove it from your video

It's basically a wrapper around whisper and ffmpeg but looks pretty wizard.

It's a lil janky bit it's 90% of the way, you may need to do some cleaning up in a proper editor.

✨ Features

• Automatic Transcription: Upload your video and get an accurate transcript with timestamps
• Intuitive Selection: Click and highlight text segments you want to remove
• Keyboard Shortcuts: Press 'X' to mark selected text for deletion
• One-Click Processing: Create and download your edited video with a single click
• Silence Detection: Automatically detects and marks silent sections in your video
• GPU Acceleration: Utilizes available hardware acceleration for faster processing

🚀 Getting Started

Prerequisites

• Python 3.8+
• Flask
• FFmpeg
• OpenAI Whisper

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/wizardcut.git
   cd wizardcut

2. Install the required packages:

   pip install -r requirements.txt

3. Install FFmpeg

figure this one out yourself, on linux you can install from package but on
windows ya gotta get it in your $PATH somehow, good luck

4. Run the application:

   python app.py

5. Open your browser and navigate to:

   http://localhost:5000
   

📋 How to Use

1. Upload a Video:

   • Drag and drop a video into the upload area, or click "Select Video File"

2. Review the Transcript:

   • After uploading, the video will be transcribed automatically
   • The transcript appears on the right side of the screen

3. Mark Segments to Remove:

   • Method 1: Click the "Select Text" button, then click on start and end words
   • Method 2: Select text with your cursor and press "X"
   • Method 3: Right-click on selected text and choose "Mark for Cutting"

4. Process Your Edits:

   • Click "Create & Download Edited Video" to process and download your edited video

5. Clear Selections:

   • To start over, click "Clear Selections"

🧙‍♂️ Technical Details

Architecture

• Backend: Flask (Python)
• Frontend: Vanilla JavaScript, HTML, CSS
• Transcription: OpenAI Whisper
• Video Processing: FFmpeg

Files and Directory Structure

• app.py: Flask server and backend logic
• static/
  • scripts.js: Frontend JavaScript functionality
  • styles.css: CSS styling
  • images/: Image assets
• templates/
  • index.html: Main HTML template
• uploads/: Temporary storage for uploaded videos
• processed/: Storage for processed videos and transcripts

API Endpoints

• GET /: Serves the main application page
• POST /upload: Handles video uploads and transcription
• POST /edit: Processes video edits
• GET /video/<session_id>/<filename>: Serves videos for playback
• GET /download/<session_id>/<filename>: Serves videos for download
• POST /cleanup: Removes old session data

🔧 Configuration

The application can be configured by modifying the following variables in app.py:

• app.config['UPLOAD_FOLDER']: Directory for uploaded files
• app.config['PROCESSED_FOLDER']: Directory for processed files
• Whisper model size: Change whisper.load_model("medium") to one of "tiny", "base", "small", "medium", or "large" to adjust the balance between transcription speed and accuracy

📝 License

GPLv3

🙏 Acknowledgements

• OpenAI Whisper for transcription
• FFmpeg for video processing
• Flask for the web framework

---

Crafted with a bit of magic ✨ by the team at SWMG Labs