Video Screenshot Extractor identifies every place in an existing blog post that references a visual claim and injects high-quality still frames extracted f…
Video Screenshot Extractor identifies every place in an existing blog post that references a visual claim and injects high-quality still frames extracted from the source video (YouTube, Vimeo, Loom, or direct MP4). It runs as a portable shell-based pipeline (Unix shell + Python 3.8+ + ffmpeg), verifies or installs dependencies, uses only relative paths under --work-dir, and avoids hard-coded system paths. The skill leverages transcripts when available to match prose to exact seconds, extracts clean frames from the raw video (no player chrome), writes SEO-optimized EXIF metadata, and updates Markdown or HTML posts by inserting images wrapped in timestamp-deeplink links (e.g., https://youtube.com/watch?v=VIDEO_ID&t=145s). Use this pipeline step when you have both a video source and an existing article and want proof images that boost SEO, dwell time, and credibility. This is a non-generative pipeline step that modifies posts with verified, first-party screenshots.
Esta página faz parte do hub OpenClaw Skills com guias de instalação, navegação por categorias e links práticos.