semantic-video-clipper

Installation
SKILL.md

Semantic Video Clipper

Overview

Segment a long video into clips by having AI analyze subtitle semantics and identify natural topic boundaries. The lightweight Python script handles parsing, timing calculation, and FFmpeg clipping based on AI-provided segment indices. Output clips and matching subtitles with filenames basename_<index>.* in same directory as the source video.

Workflow

Step 1: AI Analysis

  • Read full subtitle content (.vtt or .srt)
  • Understand semantic flow and topic transitions
  • Identify natural split points that:
    • Align with complete sentence endings
    • Occur at topic shifts (new concepts, examples, recaps)
    • Fit within duration constraints (typically 25-60 seconds)
  • Return segment ranges as cue index pairs: [(0, 12), (12, 25), (25, 40), ...]

Step 2: Python Script Execution

  • Parse subtitles into Cue objects with timing
Related skills
Installs
18
Repository
seefreed/skills
First Seen
Jan 24, 2026