GitHub radar
Explainer Video From Any Topic: Claude Code and Codex Skill
A skill for Claude Code and Codex that transforms any subject matter into an animated explainer video with audio narration, subtitles and chapters, all rendered from code using a pipeline of build agents.
anything2explainer is a skill for Claude Code and Codex that produces full-featured animated explainer videos from any subject matter. After installing this skill by cloning the repo to your skills directory and specifying your chosen topic, the agent pipeline will perform research, generate narration, synthesize TTS with word boundaries, build a storyboard with individual shots, and spawn parallel build agents that render each frame as a Remotion (React) component. The final product is a 1280×720 H.264 MP4 video with audio narration, subtitles, chapter cards and progress bar. Generating a 2–3 minute long video takes roughly 1 hour of wall-clock time, while a high-quality 3–5 minute reference video takes about 2 hours. Both English and Chinese languages are supported out of the box. At four key moments, you will be asked for input.
Why a vibe-coder should care
Bringing the power of polished animated explainer videos to any subject matter without the need for a motion graphics artist or generative video service. This process uses entirely local infrastructure, without the need for cloud rendering or subscription services. You only need Claude Code/Codex and the usual open-source tools.
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Install the skill from https://github.com/Vincentwei1021/anything2explainer: clone the repo into ~/.claude/skills/anything2explainer, install dependencies per the README (Node 18+, ffmpeg, Python packages), then make a demo explainer video on 'What are AI agents' and show me the output MP4. Ask me if you need any TTS API keys.
Developed and tested on macOS; Linux should work, Windows is untested. Node 18+, ffmpeg, and Python required — a regular MacBook will handle it. Video rendering takes 1–3 hours in the background.
Open on GitHub▌ More finds