Skip to content

Latest commit

 

History

4 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ELI5 — Picture-First Explainer Generator

English | 日本語

Explain anything to someone who knows nothing — as a single self-contained HTML file with big SVG illustrations and minimal text.

ELI5 turns any topic into a picture-first, 5-minute read: large hand-drawn-style SVG figures carry the story, words stay in the background. The output is a single HTML file with zero external requests — it works offline, on an intranet, or double-clicked from disk.

Model-agnostic by design: any LLM that can write HTML (Claude, GPT, DeepSeek, Gemini, local models) can follow the spec in SKILL.md and produce the same quality output. No Artifacts, no proprietary tools.

What you get

  • SKILL.md — the full methodology spec (procedures, language rules, HTML hard rules, pitfalls, verification checklist). Feed it to any LLM as a system prompt to generate explainers.
  • templates/eli5-template.html — the canonical output skeleton: interactive glossary, quiz, step animations, responsive layout, light/dark themes.
  • scripts/export.sh — one-command HTML → PDF/PNG export (auto-discovers Chromium).
  • examples/ — a real generated explainer in three languages: What is an API? (what-is-api-{en,ja,zh}.html) — open it to see the format in action.

Feature highlights

Feature What it does
🖼️ Picture-first Each step = a large inline SVG + one-line caption. Text is the supporting actor.
🌍 Multi-language Auto-detects zh / en / ja / ko, or one-shot generates all four from the same SVGs.
🧩 Interactive glossary Hover any term for a plain-language bubble; a glossary section recaps at the end.
Quiz recap 3–4 multiple-choice questions with instant right/wrong feedback and "go back to step N".
📦 Zero-dependency output Single self-contained HTML — no CDN, no fonts, no external JS. Offline-safe.
♻️ Model-agnostic Any LLM can reproduce the format from the spec alone.

Quick start

Option 1 — ask an AI assistant. Load SKILL.md into your agent (as a skill or system prompt), then say: eli5 <topic> or "explain X like I'm 5". The assistant writes the HTML.

Option 2 — hand-author from the template. Copy templates/eli5-template.html, fill each step with your own SVG + one caption, keep the text minimal.

Option 3 — study the example. Open examples/what-is-api-en.html in a browser to see the target quality bar.

Export to PDF/PNG (optional)

./scripts/export.sh examples/what-is-api-en.html pdf
./scripts/export.sh examples/what-is-api-en.html png

The script auto-discovers chromium / google-chrome / Playwright caches. No browser on the server? Open the HTML and Ctrl+P → Save as PDF (disable headers/footers in print settings).

Design philosophy

"If you can't explain it simply, you don't understand it well enough."

ELI5 exists because the best way to verify understanding is to explain to a beginner — with pictures. The output is designed to be read in under 5 minutes, understood without prior knowledge, and beautiful enough to share.

License

MIT

About

Explain like I am 5: picture-first single-file HTML explainers, any model, multi-language. /eli5 skill + template + export + examples.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages