Short Video Copy Extractor

The Problem It Solves

In short video operations and content creation workflows, extracting copy from trending or competitor videos is a common practice for topic research and drafting. Manually doing this requires opening links, playing videos, and transcribing or copying text, a process that is time-consuming and error-prone when dealing with volume or comparative analysis. The complexity increases further when content spans different platforms like Douyin, Kuaishou, Xiaohongshu, and WeChat Channels, each with distinct sharing link formats and content presentation. This skill's core purpose is to address this inefficient, repetitive need for copy extraction and initial processing.

How It Works

The skill functions as an automated tool that accepts a public short video share link as input and executes a series of structured processing steps:

Core Capabilities:
* Automatic Multi-Platform Adaptation: Identifies the platform by analyzing link characteristics (supports douyin, kuaishou, xiaohongshu, channels, etc.).
* Structured Data Extraction: Precisely extracts the video's title, description, and the core voiceover script.
* Compliance Pre-scan: Performs a universal rule-based check for prohibited and sensitive words on the original extracted script, providing a preliminary compliance reference.
* Copy Version Generation: Automatically generates an Optimized Reading Version (a voiceover script better suited for filming) and a Concise Version (a summary suitable for fast-paced content) based on the original script.
* Quantitative Analysis: Counts the total characters and provides a reference for voiceover duration based on average speech speed.

Key Steps and Output Structure:

Upon inputting a public share link, the skill returns a clearly formatted, structured document containing the following fixed sections:

1. 【Title】 and 【Description】: The directly extracted video metadata.
2. 【Voiceover Script · Original Extraction】: The complete, unmodified transcribed text.
3. 【Prohibited / Sensitive Word Alert】: The detection result, accompanied by a disclaimer that it is for "general reference only, not an official review."
4. 【Voiceover Script · Optimized Reading Version】: A version of the original that has been adjusted for phrasing, punctuation, and flow to better suit vocal delivery.
5. 【Voiceover Script · Concise Version】: A distilled summary that extracts core viewpoints and removes redundancy.
6. 【Script Duration Reference】: Provides a word count and an estimated duration.

The entire output is cleanly formatted and ready for direct use in notes, scriptwriting, or data analysis.

Scope and Limitations

• Data Dependency: The skill processes data from publicly accessible videos. Extraction will likely fail for videos with access restrictions or that require a login.
• Limitations of Compliance Alerts: The prohibited/sensitive word detection is based on a general rule set. Results are for reference only and cannot substitute for the official content review standards or real-time policies of each platform. Manual judgment is still required before final publication.
• Voiceover Duration Estimate: The duration estimate is based on an average speech rate (approx. 4 characters/second). Actual recording time varies by individual speaking habits and pauses, so it should be adjusted based on practical context.
• Output Focus: The skill is focused on copy extraction and basic optimization. It does not provide video downloading, deep data analytics, or interactions related to a user's account.