Description

Organize short video links, spoken scripts, or image text into analyzable content profiles while restoring the original text, platform, word count, speaking rate, and estimated duration. Whether for Douyin, Kuaishou, Xiaohongshu, Channels, Bilibili, or YouTube, it identifies e-commerce, science popularization, storytelling, tutorial, emotional, or paid knowledge types based on specific signals in the script. When the original text is unavailable, it explicitly states so without guessing based on titles or covers. On this basis, the Skill breaks down short video structures across five stages: hook, setup, core, pivot, and CTA. It diagnoses the function of each segment, the performance of key seconds, and structural order. It scores dimensions such as hook strength, information density, emotional tension, and action guidance, providing clear score deduction reasons and actionable improvement steps. Compliance screening flags risks such as absolute expressions, false promises, earnings and efficacy claims, and private traffic redirection, while offering directly replaceable wording. When script revision is needed, it generates 15-second, 30-second, and 60-second spoken versions tailored to different video durations based on the original text, retaining core information and explaining compression trade-offs without inventing facts or promises. It supports batch comparison of multiple scripts, horizontal analysis of competing or reference drafts, and delivery in Markdown, JSON, CSV, and plain text. This makes it suitable for content operations, directors, media buyers, and independent creators to review hit content, conduct compliance checks, and revise scripts before shooting.