clips-machine
mayank8290 作
Coding Agents & IDEs概要
Transform long videos into viral short-form clips.
説明
このスキルでできること
Clips Machine takes long videos from YouTube, local files, TikTok, Twitter/X, or Twitch and cuts them into short vertical clips ready for TikTok, Reels, or Shorts. It transcribes audio locally with Whisper, scores each moment by virality criteria, and exports captioned 9:16 clips with multiple text style options.
Runs entirely on local tools with no cloud service, so there are no per-clip fees, upload limits, or API keys required.
使いどころ
- Clipping the top 5 moments from a 2-hour podcast
- Turning a Twitch gaming VOD into meme-captioned TikToks
- Extracting quotable segments from a conference keynote on YouTube
- Processing a local interview MP4 into Reels-ready highlights
- Batch-generating 10 short clips from a long YouTube interview for a client
ワークフロー例
User asks: Take this podcast and find the 10 best moments: https://youtube.com/watch?v=PODCAST_ID — make them Hormozi-style with bold captions
- 1
Downloads the video using yt-dlp
- 2
Transcribes the audio locally with Whisper.cpp to produce a timestamped transcript
- 3
Analyzes the transcript and scores moments by virality criteria (hooks, emotional peaks, quotable lines)
- 4
Cuts the top 10 segments using FFmpeg and crops to 9:16 vertical format
- 5
Renders animated Hormozi-style captions onto each clip and saves them to ~/Videos/OpenClaw/
10 captioned vertical MP4 clips with a summary.md listing each clip's virality score and timestamp
要件
- ffmpeg (brew install ffmpeg or apt install ffmpeg)
- yt-dlp (brew install yt-dlp or pip install yt-dlp)
- whisper-cpp (brew install whisper-cpp or built from source)
インストール
clawhub install mayank8290/clips-machineCLIコマンドを実行するか、Claude Code、Cursor、Codexなどお使いのコーディングエージェントにプロンプトを貼り付けてください。