How It Works

The script uses ffmpeg's scene detection to identify visual cuts in video files. When you first use it on a video, it processes a small section around the cursor to find nearby cuts. This happens quickly and does not require processing the entire video. A background processor then continues caching the rest of the video so future jumps are instant.

Key Features

• Incremental processing: Only processes what's needed for immediate navigation, then fills in the rest in the background.
• Smart video detection: Supports videos in subprojects/master projects. If your project has only one video, the script uses it automatically. If multiple videos exist, position your cursor on the video item you want to navigate.
• Persistent cache: Cut positions are saved with your project, so reopening a project doesn't require reprocessing.

Configuration

Advanced users can modify config.lua in the script folder to adjust:

• SCENE_THRESHOLD — Detection sensitivity (default: 0.1, lower = more sensitive)
• PROCESSING_WINDOW_SIZE — How much video to cache around cursor (default: 480 seconds / 8 minutes)
• ENABLE_PROCESSING_WINDOW — Set to false to cache entire video instead of just around cursor
• MAX_SEARCH_DISTANCE — How far to search before giving up (default: 120 seconds)