Motion Video Kit is a Claude Code skill for creating premium, launch-style commercials for businesses using AI-generated footage, code-built motion with HTML and GSAP, and selective Three.js rendering.
The tool addresses the challenge of producing high-quality commercial videos at scale by packaging patterns and processes learned from analyzing professional SaaS launch films and building two complete sample commercials through iterative critique. Its core approach centers on an independent-critic loop where the builder and judge are separate roles, applying fresh evaluation to actual renders with item-by-item verification tracked in a ledger. The kit provides ready-to-use critic prompts, a motion grammar of six rules and sixteen reusable mechanisms, measurable quality checks covering frozen time, loudness, contrast and brand color, audio rules for music and sound effects, and a business playbook with verticals, pricing anchors, and offer templates. For Three.js work, it includes patterns for deterministic seekable scenes, exploded layers with projected callouts, photos pinned in 3D space, and techniques for matching real device motion frame by frame.
Adoption suits teams building commercials for SaaS and product launches who want systematic quality control and reusable motion patterns rather than starting from scratch. The kit works best with HyperFrames for rendering but applies its principles, prompts and checks across any renderer. The principles are also usable as plain context for any LLM beyond Claude Code. The repository contains original written analysis and links to reference films rather than redistributed footage, and includes scripts requiring ffmpeg and ffprobe with the ebur128 filter, with optional Node and HyperFrames support for the lab template.
Development shows sustained focus on a narrow, well-defined problem. The project includes detailed reference notes on all twenty-eight analyzed films with links to originals, measurable quality criteria grounded in client enforcement practices, and specific technical solutions like deterministic accumulation motion blur and locked-camera techniques for product realism. Documentation covers both the conceptual motion grammar and concrete implementation details across scripts, templates and patterns. The kit enforces labeling to prevent presenting AI-generated imagery as real customer results.
