Behind the scenes

One call. Three pages.

This is the request the agent sent. The HTML and CSS are trimmed here, but nothing else is.

# POST https://wordy.page/api/agent
{
  "action": "publish",
  "slug":   "agent",
  "title":  "Agent Mode Showcase",
  "css":    ":root{--hot:#ff5c7a;--mint:#5cf2c4} .hero{background:photo(UZOpP-YHe9Q)...}",
  "pages": [
    { "path": "",        "html": "<header class=hero>..." },
    { "path": "gallery", "html": "<div class=gal>..." },
    { "path": "made",    "html": "<div class=term>..." }
  ]
}
1API request
3pages
0lines of JavaScript
4library photos
Editor vs. agent mode

Same platform. Different ceiling.

CapabilityHuman editorAgent mode
LayoutBuilt-in page templateAny HTML structure
StylingTheme colors and fontsA full custom stylesheet
AnimationNoneKeyframes and transitions
PagesOneUp to ten
PhotosPick from libraryLibrary photos placed anywhere in HTML or CSS
SafetySanitizedSanitized: no scripts, external URLs, or imports
Connect

Bring your agent.

Use a personal API key, or connect any MCP client with OAuth at https://wordy.page/mcp/auth. Your agent signs in to your account and publishes as you.

Your agent can build this too.

Hand it a Wordy Page key or connect it over MCP. It writes the HTML and CSS, and the site goes live.

Read the agent guide