Skill Detail

Live Stream Clip Extractor

Extracts highlight clips from live streams using Twitch Helix API, YouTube Live Streaming API, and FFmpeg segment detection. Identifies peak moments via chat velocity analysis and audio energy spikes.

Media & TranscriptionChatGPT Agents
Media & Transcription ChatGPT Agents Published
Tool match: ffmpeg โญ 59.4k GitHub stars
INSTALL WITH ANY AGENT
npx skills add agentskillexchange/skills --skill live-stream-clip-extractor-agent Copy
Works best when you want a reusable capability, not another fragile one-off prompt.
At a glance
Last updated
Mar 24, 2026
Quick brief

The Live Stream Clip Extractor automatically identifies and extracts highlight moments from live streaming content. It monitors live streams via the Twitch Helix API clips endpoint and YouTube Live Streaming API liveBroadcasts resource. The agent detects peak moments using multiple signals: chat message velocity analysis via Twitch IRC (TMI.js), audio energy spike detection using FFmpeg loudnorm filter and ebur128 measurement, and scene change detection via FFmpeg scene filter with configurable thresholds. Clip extraction uses FFmpeg segment muxer for frame-accurate cutting without re-encoding, and the agent generates multiple output formats (MP4, WebM, GIF) with automatic thumbnail generation at the highest-energy frame. It integrates with Streamlink for reliable stream capture across platforms and uses yt-dlp for VOD segment downloading. The extractor supports automated posting to social media via the Twitter API v2 media upload endpoint and Instagram Graph API for Reels. Clip metadata includes timestamp, duration, chat context, and engagement metrics. Batch processing handles full VOD analysis with configurable highlight density and minimum clip duration settings.