Nanopage
← Dashboard
Developer

MCP interface

Connect your AI coding agent (Claude Code, Claude Desktop, Cursor) to nanopage. Your agent generates the site using its own tokens — nanopage validates, stores, and hosts it on a nanopage.site subdomain.

New to it? Read the MCP server overview →

How it works
1

Add the endpoint

Point Claude Code, Claude Desktop, or Cursor at the server URL using the snippets below.

2

Approve access

Your client opens the browser: sign in to nanopage and approve. No token to copy — CI and scripts can still use a personal access token.

3

Publish sites

Ask your agent to build and publish. Your site goes live at {slug}.nanopage.site in seconds.

Server endpoint
https://api.nanopage.site/mcp
Personal access tokens0 active

Loading your tokens…

Set up your client

Sign-in happens in your browser

Nanopage speaks standard MCP OAuth: add the endpoint URL and your client discovers the rest — it registers itself, opens your browser to sign in, and asks you to approve access. Nothing to paste. Personal access tokens (above) remain the right tool for CI, scripts, and clients that cannot open a browser — see the Token (CI / scripts) tab.

Adds the server to your user-level Claude Code config. On first use Claude Code opens your browser to sign in to nanopage and approve access — no token to paste.

terminal
claude mcp add --transport http nanopage https://api.nanopage.site/mcp

Restart the client after adding it. Verify with /mcp inside Claude Code — you should see nanopage listed as connected.

What the server offers

Tools

  • publish_websitewrite

    Create a new nanopage site from a file bundle (index.html + assets + images).

  • publish_versionwrite

    Add a new version to an existing site. Auto-promotes to current unless promote:false.

  • create_uploadwrite

    Upload URLs for images and other binaries, so the bytes never have to ride inside a tool call.

  • list_websitesread

    List every website in your account.

  • get_websiteread

    Full details for one website: current version, past versions, files.

  • list_versionsread

    All versions of a website, newest first.

  • get_job_statusread

    Look up a job. Publish jobs land ready immediately.

  • suggest_slugread

    Check slug validity + availability and get a fallback suggestion.

Resources

  • nanopage://spec/authoring-guide.md

    The authoring spec: file layout, hard rules, size limits. Your agent should read this before generating.

  • nanopage://websites/{websiteId}/versions/{versionId}/{path}

    Read a specific file from a stored version.

Prompts

  • nanopage_publish_site

    Bootstraps a nanopage-publishable site from a topic. Inlines the spec so the model sees the constraints before it generates.

Rules your agent must follow

Nanopage hosts static sites. The following are rejected at publish time:

  • • External <script src>, <link href>, or fonts (no CDNs — bundle everything inline).
  • • <iframe>, <object>, or <embed>.
  • • javascript: URLs, service workers, off-site meta-refresh redirects.
  • • Forms submitting to external hosts.
  • • Anything expecting a backend (no server APIs, WebSockets, or databases).

Allowed layout: index.html + up to 14 more {slug}.html, assets/site.css, assets/{slug}.js, images/, files/, videos/. Limits: 50 MB per file, 1 GB per site, no file-count limit.

Dynamic content: add collections.json + templates/{collection}/list.html (optionally item.html, templates/pages/{page}.html) and Nanopage creates the collections, seeds them, and renders the pages — blog, news, products, events, galleries, contact forms — which the owner then edits in the editor without a republish. The spec resource documents the manifest and template language.