skills.rest
lcbuaaliu
lcbuaaliuCOMMUNITY
Explore all skills by this author →
GitHub

AI剪口播

Identifies speech errors, filler words, repeated phrases, and silence in Chinese talking-head videos for editing workflows.

446|57|Updated Jun 9, 2026
Search Tags:#audio transcription#video editing#volcengine#subtitle generation#fcpxml#speech cleanup#waveform review
One-click install
npx skills add https://github.com/lcbuaaliu/ai-jian-koubo --skill ai-lcbuaaliu
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: AI剪口播
Source: https://github.com/lcbuaaliu/ai-jian-koubo/tree/main
Command: npx skills add https://github.com/lcbuaaliu/ai-jian-koubo --skill ai-lcbuaaliu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires bash, node, python3, ffmpeg, ffprobe, curl, volcengine speech api, and includes scripts (resource) and assets (resource) components.

What problem does it solve?

This Skill reduces the time and effort required to edit talking-head videos by identifying repeated speech, filler words, stutters, incomplete phrases, and unnecessary silence before final editing.

Core Features & Use Cases

  • AI-Assisted Speech Cleanup: Transcribes videos and preselects verbal mistakes, filler words, repeated phrases, and silent sections for deletion.
  • Waveform Review: Launches a local browser-based review interface with transcript controls, audio waveform visualization, video preview, and adjustable silence padding.
  • Editing Workflow Export: Generates FCPXML projects that can be imported into Jianying or Final Cut Pro, while optionally recording confirmed edits for personalized rule learning.
  • Use Case: A creator can submit a Chinese talking-head video, review the AI-selected cuts locally, export the confirmed timeline as FCPXML, and finish the edit in their preferred video editor.

Quick Start

Ask the coding agent to process a talking-head video, choose either speech cleanup or subtitle generation, and follow the local review workflow to export the finished result.

Frequently Asked Questions about AI剪口播

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automatically identify filler words and speech errors in talking-head videos?▼

AI-assisted speech cleanup transcribes talking-head videos using the Volcengine Speech API and preselects verbal mistakes, filler words, repeated phrases, and silent sections for deletion before final editing.

Can I export an FCPXML project with AI-preselected cuts for Final Cut Pro or Jianying?▼

You can export an FCPXML project containing confirmed timeline edits, which is directly importable into Final Cut Pro or Jianying to finish the video editing workflow.

Do I need Node.js and Python 3 to run local waveform review for video transcription?▼

Yes, local waveform review and video transcription require Node.js, Python 3, FFmpeg, curl, and a Volcengine Speech API key to launch the browser-based review interface.

What is the best way to review and adjust silence padding in Chinese video transcription?▼

The best way is using the local browser-based review interface, which provides audio waveform visualization, transcript controls, video preview, and adjustable silence padding for Chinese spoken content.

Does this speech cleanup workflow support personalized rule learning for future edits?▼

The speech cleanup workflow optionally records your confirmed edits during the local review process, enabling personalized rule learning to improve future transcription and editing automation.

Related Skills

Explore More Design & Creative Skills→
Quriosity-agentQuriosity-agent
35

videocut:talk-edit

Transcribe talking-head videos and detect speech errors for review.

Official
Advanced
MadAppGangMadAppGang
7

final-cut-pro

Generate FCPXML project structures for Final Cut Pro 10.4+.

Official
Intermediate
abdu6666abdu6666
17

speech-transcriber

Convert audio to text with timestamps using Whisper, generating SRT and JSON outputs.

Community
Intermediate
aegnticaegntic
4

clip-hand-skill

Cuts video into short clips with burned-in captions from uploaded files or URLs.

Community
Advanced