Turn a Sentence Into an Interactive Diagram With Archify

You've got a system in your head—a request flow, a travel plan, a learning path—and the fastest way to explain it is still a whiteboard or a messy sketch. What if you could just describe it to your AI agent and get back something you can actually click through? That's the pitch behind Archify, an open-source agent skill that turns plain-language descriptions into interactive HTML visuals.

What It Does

Archify is an agent skill that converts a natural-language description into an interactive HTML diagram. You describe what you want to understand, plan, or share—a web request flow, a travel itinerary, a learning map, a complex system—and Archify generates visual output you can explore, customize, and share. The project's own example is refreshingly concrete: "Browser calls the API, the API checks Redis, and a cache miss queries PostgreSQL." That's the level of input it's designed for.

It works as a skill for AI coding agents, with documented support for Cursor, Claude Code, Codex CLI, and OpenCode, plus additional integrations available through the installation options. The output is interactive HTML, which means the diagrams aren't static images—you can follow source links and trace paths through them. The current stable version is 3.0.1, and it's MIT licensed.

Why It's Cool

• It meets you where you already are. You're probably already talking to an AI agent about your code or your plans. Archify doesn't ask you to learn a diagramming tool, a DSL, or a new file format—it plugs into the agent workflow you've got and turns your existing description into something visual.

• The output is interactive, not a screenshot. Static diagram exports are where good explanations go to die. Because Archify produces HTML, the result stays explorable. The README highlights following source links and tracing a path through the diagram, which is exactly what you want when you're walking someone through a system.

• It's genuinely broad in scope. The README doesn't pigeonhole this as a software architecture tool. Travel itineraries and learning maps sit alongside complex systems and repo diagrams. That flexibility is the interesting part—the same mechanism that maps a cache-miss path can map a trip.

• Distribution is a single command. npx skills add tt-a1i/archify -g is about as low-friction as installs get. No clone, no build step, no config file archaeology.

• Multiple language READMEs and a real community. There's Simplified Chinese and Japanese documentation, plus Discord, WeChat, and QQ communities. That's a signal the project has actual users, not just stars.

How to Try It

1. Install the skill globally:

npx skills add tt-a1i/archify -g

1. Open your agent (Cursor, Claude Code, Codex CLI, or OpenCode) and describe what you want visualized. The README's example prompt:

Use Archify to diagram a web request: Browser calls the API,
the API checks Redis, and a cache miss queries PostgreS

1. Explore the generated interactive HTML—click through the diagram, follow source links, and trace paths.

If you'd rather see it before installing, the project hosts a gallery of interactive examples you can poke at directly, along with a scenario guide for ideas on what to build. There's also a 35-second demo video in the README showing a repo being mapped from a single sentence.

Everything lives at github.com/tt-a1i/archify.

Final Thoughts

Archify is a focused tool that does one thing: turning your description into a visual you can interact with. It's not trying to be a full diagramming suite, and it doesn't need to be. If you regularly explain systems, plans, or processes to other people—or just to future-you—and you already work inside an AI agent, this is a low-cost addition to that workflow. The single-command install and the interactive HTML output are the two details that make it worth a try rather than just a bookmark. Give it one real diagram from your own work and see whether the output saves you a whiteboard session.