# anything2explainer

A Claude Code and Codex skill that turns a topic or document into a narrated, code-animated explainer video, with TTS voiceover, word-aligned subtitles and a chapter progress bar, using Remotion rather than stock footage or a video model.

anything2explainer is an agent skill that takes a topic, or an article to turn into a video, and produces a finished narrated explainer film with no stock footage, generative video model, or frames borrowed from other work. Every frame is drawn in code with Remotion, a React and TypeScript video framework, and the output is a 1280x720 MP4 with synchronized text-to-speech voiceover, word-boundary-aligned subtitles, chapter cards, a heads-up display and a bottom chapter progress bar, in either Chinese or English.

What ships is the whole method an agent needs to finish a film rather than a single command: a compilable Remotion template, a library of visual primitives and lighting, tooling for voiceover generation, storyboarding, rendering and quantitative quality checks, written style and motion specifications, a protocol for dividing the work across multiple parallel build agents, and one complete reference film as the quality bar. The agent researches the topic with sources, writes the narration, generates a frame-accurate voiceover timeline, storyboards every shot, dispatches parallel agents that each write one Remotion component per shot, and runs separate QC agents against the rendered frames before delivery, typically taking one to three hours of wall-clock time.

It is written in TypeScript.

Website: https://github.com/Vincentwei1021/anything2explainer