Short Video Voiceover Script Factory

Problem

Short-video scripts often read like copy, not like lines a solo creator can perform on a phone camera. Source material may be an opinion, experience, or SOP without a hook, pacing, or clear visual boundaries. Reading it directly can sound written, and forced viral tactics may invent data, plot twists, or efficacy claims.

How It Works

The skill is built for creators with roughly 0–100k followers who produce selfie-style talking videos. It converts text material into a script that can be read almost verbatim:

• Material diagnosis: identify the main type (story, product, knowledge, opinion, etc.), core points, and target length; weak material gets a suggestion to add concrete experiences or data.
• Hook and structure: prioritize conflict, contrast, or pain points found in the source instead of forcing templates; opinion-led pieces use a resonance bridge before argument.
• Oral pacing: keep sentences short, mark 【tone】 and 【pause】, and adjust information density for 30s, 60s, or 2–3min versions.
• Compliance and fact limits: avoid inventing numbers, amounts, times, or follow-up events; soften sensitive efficacy claims while preserving core facts; anonymize private details during drafting.
• Output and iteration: return a concise spoken script by default; when a full version is requested, add selfie/B-roll/text-overlay suggestions, pinned-comment guidance, and publishing tips, with compliance notes grouped at the end under a pre-publish checklist.

Boundaries

It is intended for turning existing opinions, experiences, knowledge, product feedback, or hot-topic comments into phone-selfie-ready voiceover scripts. It is not designed for fictional stories, multi-camera dramas, pure video generation, live-stream scripts, or long-video storyboards. The output is a script, not a guarantee of viral performance.