SuperImg Skill

Mental Model

Video is a pure function of time. Your template's render(ctx) is called once per frame. It receives a RenderContext and returns an HTML string. All animation is derived from sceneProgress (0→1 over the scene) or sceneTimeSeconds (elapsed seconds). Use std.tween as the canonical animation primitive. Data comes from data, merged with per-scene overrides. No timeline editor, no keyframes — just math and HTML.

Quick Start

import { defineScene } from "superimg";