Colorfigures

Describe an everyday object and AI draws it as a hand-inked isometric model with working buttons that you can turn around and see inside.

Ecosystem: Agent Skills
Skill category: Design
Creator / organization: April Zhu（X：@Aprilzz423）

View work and source ↗

What it is

Colorfigures is a Skill by April Zhu. Its site puts it this way: describe an everyday object, and AI draws it hand-inked and isometric, with buttons that actually work, and you can turn it all the way round and see inside. The result is a single HTML file you can rotate, look through and operate. The site shows 26 examples. The author notes it isn't suited to people, animals or plants.

Who it's for:Creators and teachers who want to explain how everyday devices work and need interactive illustrations.

Before you start: The Skill is free and public, and the site says it's under the MIT licence; your Claude, ChatGPT, Gemini or Codex plan or API costs are separate. The site says the author's own tests took about 15 to 30 minutes and 200k to 400k tokens per object. The install command downloads an archive with curl and unpacks it straight into your Skill folder; we didn't run it. The site says the generated HTML has no tracking, needs no account and makes no outside requests; that's the author's statement, which we haven't checked file by file. The site is in English, and it doesn't say how well Chinese descriptions work.

Based on the official site's public information. Not tested by AI Inspiration World.

Based on the project's public documentation. Not tested by AI Inspiration World.