clip-selection

What problem does it solve?

Long-form video content often lacks ready-to-share short-form moments. This Skill identifies, scores, and extracts cohesive 45-90 second clips from transcripts, enabling efficient social-first video production.

Core Features & Use Cases

• Transcript-driven parsing: convert word-level transcripts into structured clip candidates with precise timestamps.
• Anchor-word based clipping: anchor exact verbatim phrases (5-8 words) to ensure precise, non-mid-sentence boundaries.
• Scoring & ranking: apply the 5-category framework to produce clip metadata, captions, and a ranked selection for batch reframing.

Quick Start

Analyze a transcript file and run through parse, select, score, and output steps to generate clip definitions and a report.