A free skill that teaches Claude Code or Codex to draw proper Excalidraw diagrams — flowcharts, wireframes, slide decks, business canvases — right in your own Excalidraw workspace.
You type a sentence like one of those. Your AI assistant opens your Excalidraw+ workspace, draws the diagram in Excalidraw's friendly hand-drawn style, checks its own work with a screenshot, and hands you a link. The drawing is fully editable — move boxes, change words, share it with your team, just like any whiteboard you made yourself.
No design tools to learn. No dragging boxes around. You describe; it draws.
Step 1 — Get your Excalidraw key. Sign in at app.excalidraw.com (Excalidraw+), open Settings → API keys, and create a key. Copy it somewhere safe — treat it like a password.
Step 2 — Add the skill. Paste this into your terminal (the same place you run Claude Code or Codex):
npx skills add NicholasSpisak/excalidrawThat's the skills installer — it sets things up for Claude Code, Codex, and most other AI coding assistants in one go.
Step 3 — Tell it your key. Run these two lines once, then restart your terminal:
mkdir -p ~/.config/excalidraw && printf '%s\n' 'YOUR-KEY-HERE' > ~/.config/excalidraw/api_keyYour key stays in one private file on your computer — never in any project or code you share.
The skill works out of the box. If you prefer wiring the official Excalidraw MCP server directly into your assistant:
Claude Code — add to ~/.claude.json:
{
"mcpServers": {
"excalidraw": {
"type": "http",
"url": "https://api.excalidraw.com/api/v1/mcp",
"headers": { "Authorization": "Bearer <API_KEY>" }
}
}
}
Codex — add to ~/.codex/config.toml:
[mcp_servers.excalidraw]
url = "https://api.excalidraw.com/api/v1/mcp"
env_http_headers = { Authorization = "EXCALIDRAW_API_KEY" }
Full details in the README on GitHub.
All seventeen of Excalidraw's official use cases, each with its own layout playbook baked in:
Every drawing lands in your own workspace, gets screenshot-checked before it's called done, and comes back as a shareable link.
This skill is one small example of what happens when business owners get practical with AI. That's the whole idea behind AI Operator Academy — a community where owners learn to turn AI into real systems (and real revenue), no engineering degree required.
Take a look at the Academy →