What problem does it solve? Creating rendered videos from text, scripts, or websites normally requires video editing software and manual timeline work. This Skill turns HTML/CSS/JS compositions into deterministic MP4/WebM videos by capturing pages frame-by-frame and encoding them with FFmpeg. ## Core Features & Use Cases - HTML-to-Video Rendering: Author compositions as HTML files with data-* timing attributes and GSAP timelines, then render to MP4 or WebM at 24/30/60fps in draft, standard, or high quality. - Captions, TTS, and Audio-Reactive Visuals: Generate local Kokoro TTS narration in 8 languages, transcribe with Whisper for word-level captions, and drive visuals from pre-extracted audio frequency bands. - Website-to-Video Pipeline: Capture a URL into screenshots, assets, and design tokens, then follow a 7-step workflow (capture, DESIGN.md, SCRIPT.md, storyboard, composition, render, deliver) to produce a promo video. - Use Case: A user provides a product landing page URL and asks for a 30-second social ad. The Skill captures the site, extracts its brand colors and fonts into a DESIGN.md, writes a narration script, builds a multi-scene composition with shader transitions, and renders a final MP4. ## Quick Start Ask the agent to render a short animated promo video from a script or website URL, and it will scaffold a project with npx hyperframes init, build the composition, and render it to an MP4 file.
