Developers

Nanopage Developer Resources

Everything machine-readable and agent-facing that Nanopage publishes: the REST API and its OpenAPI spec, the MCP server, llms.txt, sitemap, and markdown content negotiation.

Nanopage MCP server

Nanopage exposes its publishing API as a Model Context Protocol server, so Claude Code, Claude Desktop, Cursor, Codex, and other AI agents can build and publish hosted websites natively.

https://api.nanopage.site/mcpStreamable HTTP · JSON-RPC 2.0 · bearer-token auth
  • publish_website

    Create a new site from a file bundle

  • publish_version

    Add a version to an existing site

  • list_websites

    List every website in the account

  • get_website

    Full details for one website

  • list_versions

    All versions, newest first

  • get_job_status

    Look up a publish job

  • suggest_slug

    Slug validity + availability

Mint a personal access token and grab client config snippets on the MCP interface page. Read the announcement post for a walkthrough.

Nanopage REST API

The same seven operations as the MCP server, as plain HTTPS + JSON — for agents and scripts that speak REST instead of MCP. The full surface is described by an OpenAPI 3.1 spec at nanopage.site/openapi.json (also served at api.nanopage.site/openapi.json).

https://api.nanopage.site/v1JSON over HTTPS · bearer-token auth · same tokens as MCP
  • listWebsites

    GET /v1/websites

  • createWebsite

    POST /v1/websites

  • getWebsite

    GET /v1/websites/{websiteId}

  • listWebsiteVersions

    GET /v1/websites/{websiteId}/versions

  • createWebsiteVersion

    POST /v1/websites/{websiteId}/versions

  • getJob

    GET /v1/jobs/{jobId}

  • suggestSlug

    GET /v1/slugs/suggest?value=…

Scopes: every personal access token carries the single scope mcp, which grants exactly these operations plus the MCP endpoint — never the dashboard, billing, domains, or admin. The scope list 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="…". Mint tokens on the MCP interface page.

Machine-readable site
  • openapi.json

    OpenAPI 3.1 spec for the Nanopage REST API — every operation with typed schemas and operation IDs.

  • llms.txt

    Compact navigation index of the product and site, written for LLMs and agents.

  • llms-full.txt

    Complete annotated index: every use-case page, template, and blog post with descriptions.

  • sitemap.xml

    Every indexable page with last-modified dates.

  • robots.txt

    AI crawlers (GPTBot, ClaudeBot, Google-Extended, PerplexityBot…) are explicitly allowed.

  • oauth-protected-resource

    RFC 9728 protected-resource metadata declaring the supported API token scopes.

Markdown content negotiation

Core pages serve text/markdown via the Accept header (with Vary: Accept), and at explicit .md URLs. 404s return a short markdown body pointing agents at the sitemap and llms.txt. 

curl -H "Accept: text/markdown" https://nanopage.site/
/index.md/pricing.md/how-it-works.md/faq.md/about.md/vs.md/use-cases.md/templates.md/blog.md/contact.md/developers.md/mcp.md

Questions about the MCP server or agent access? Email hello@nanopage.site — every email is read.