Nanopage Developer Resources

Tokens: create personal access tokens on the API tokens page and choose what each one reaches: the REST API, the MCP server, or both. No token ever reaches the dashboard, billing, domains, or admin. OAuth sign-ins from MCP clients reach both. The OAuth scope is published machine-readably as RFC 9728 protected-resource metadata at api.nanopage.site/.well-known/oauth-protected-resource, and the API's 401 responses point at it via WWW-Authenticate: Bearer resource_metadata="…".

• The slug is the identity. Sending an entry whose slug exists updates it, so you can run the same import again after fixing something. Keep your old slugs (up to 80 characters) and the URLs stay the same, at /{collection}-{slug}.
• Dates come along. publishedAt and createdAt keep your archive in order.
• HTML and Markdown work. A rich-text field takes { "html": "…" } or { "markdown": "…" } and ends up exactly as if you had pasted it into the editor: headings, lists, quotes, links, images with captions.
• Images by URL. Image fields and <img> tags in the HTML are downloaded into the collection's library, once per URL. baseUrl resolves relative paths.
• Nothing is all-or-nothing. Up to 100 entries per request, each with its own result. A collection can hold any number of entries, and every change is live on the next page view.

curl -X POST https://api.nanopage.site/v1/websites/$SITE/collections/blog/entries/batch \
  -H "Authorization: Bearer $NANOPAGE_TOKEN" -H "content-type: application/json" \
  -d '{
    "baseUrl": "https://old-blog.example.com/",
    "entries": [{
      "slug": "how-we-roast-our-coffee",
      "publishedAt": "2021-03-04T10:00:00Z",
      "data": {
        "title": "How we roast our coffee",
        "cover": "https://old-blog.example.com/uploads/roast.jpg",
        "body": { "html": "<h2>Beans</h2><p>We start early…</p>" }
      }
    }]
  }'

AI agents can do the same through the MCP tool upsert_entries: ask Claude to move your WordPress blog into Nanopage and point it at the export.