Lolipop MCP Server

Official remote MCP server for Lolipop Rental Server by GMO Pepabo. Manage domains, subdomains, free SSL, and WordPress via natural language.

GitHub

Documentation

Lolipop MCP Server

日本語版 README はこちら / Japanese README

The official remote Model Context Protocol (MCP) server for Lolipop Rental Server — a shared web hosting service operated by GMO Pepabo, Inc.

Manage domains, subdomains, free SSL certificates, and WordPress installations — all through natural language conversations with AI assistants.

Endpoint

https://lolipop.jp/api/v1/mcp

Transport: Streamable HTTP. Authentication: Personal Access Token (PAT) via the Authorization: Bearer header.

Features

  • Domain management — add, list, update (public folder), and remove custom domains
  • Subdomain management — add, list, and remove subdomains
  • SSL — list free SSL certificates and request issuance (Let's Encrypt)
  • WordPress — list installations and run a one-click install
  • Account info — retrieve the authenticated API key's account information (plan, server, rate limit)

Pricing: MCP server access is free. A valid Lolipop Rental Server contract is required.

Authentication

The Lolipop MCP server authenticates with a Personal Access Token (PAT), issued from the Lolipop user panel.

  1. Sign in to the Lolipop user panel and issue an API key (PAT). Copy the token (lp_pat_...).
  2. Register the endpoint URL together with the Authorization: Bearer <PAT> header in your MCP client (see Quick Start below).

Scopes are selected at issuance time (domains:read / domains:write / ssl:read / ssl:write / wordpress:read / wordpress:write). Requests exceeding the token's scope return 403 insufficient_scope.

Note: claude.ai (the web app's custom connectors) currently requires OAuth and cannot pass a static Authorization header, so it is not supported at this time. Use a client that supports custom headers (Claude Code, Cursor, Gemini CLI, OpenAI Codex), or the mcp-remote stdio bridge below.

Quick Start

Keep your PAT out of committed files and shell history. Store it in an environment variable (e.g. LOLIPOP_PAT) and reference it, rather than pasting the raw token into a config file or a command line. The examples below use ${LOLIPOP_PAT}.

export LOLIPOP_PAT="lp_pat_..."   # your issued token

Claude Code

claude mcp add --transport http lolipop https://lolipop.jp/api/v1/mcp \
  --header "Authorization: Bearer ${LOLIPOP_PAT}"

Claude Code MCP docs

Cursor

Add to .cursor/mcp.json. Cursor expands ${env:VAR} references, so the raw token stays in your environment, not the file:

{
  "mcpServers": {
    "lolipop": {
      "url": "https://lolipop.jp/api/v1/mcp",
      "headers": {
        "Authorization": "Bearer ${env:LOLIPOP_PAT}"
      }
    }
  }
}

Gemini CLI

Add to ~/.gemini/settings.json. Gemini CLI expands $VAR / ${VAR} references from the environment:

{
  "mcpServers": {
    "lolipop": {
      "httpUrl": "https://lolipop.jp/api/v1/mcp",
      "headers": {
        "Authorization": "Bearer ${LOLIPOP_PAT}"
      }
    }
  }
}

See the Gemini CLI MCP server documentation for details.

OpenAI Codex CLI

Codex supports remote MCP servers and can read the token from an environment variable via bearer_token_env_var. Configure https://lolipop.jp/api/v1/mcp per the Codex MCP documentation:

[mcp_servers.lolipop]
url = "https://lolipop.jp/api/v1/mcp"
bearer_token_env_var = "LOLIPOP_PAT"

Local stdio bridge (for clients without custom-header support)

Use mcp-remote to bridge a stdio-only client to the remote endpoint. Reference the token via environment variable so it does not appear in the config file:

{
  "mcpServers": {
    "lolipop": {
      "command": "npx",
      "args": [
        "-y", "mcp-remote", "https://lolipop.jp/api/v1/mcp",
        "--header", "Authorization: Bearer ${LOLIPOP_PAT}"
      ]
    }
  }
}

Documentation

Requirements

  • A Lolipop Rental Server account on a supported plan
  • A Personal Access Token (PAT) issued from the user panel
  • An MCP-compatible client that supports custom request headers (Claude Code, Cursor, Gemini CLI, OpenAI Codex CLI, etc.)

Support

License

MIT