Magic Meal Kits MCP Server Version Endpoint
Retrieve Magic Meal Kits MCP server version details to help AI assistants access up-to-date server information.
npx -y @pureugong/mmk-mcpOverview
The Magic Meal Kits MCP Server implements the Model Context Protocol (MCP) to expose version information from a Magic Meal Kits API as a tool that AI assistants can call. Instead of hard-coding backend details into an assistant, this MCP server provides a small, secure bridge so LLM clients can programmatically ask “what version is the backend running?” and get a structured answer.
This is useful for debugging, monitoring, and context-aware assistant behavior. For example, an assistant can detect when the API is running an older release and adapt responses or instruct a user to update. The server is lightweight, requires only an API key to authenticate to the Magic Meal Kits backend, and is designed to be run locally or as part of an assistant integration (e.g., Claude Desktop).
GitHub: https://github.com/pureugong/mmk-mcp
Features
- Exposes a single MCP tool to retrieve Magic Meal Kits server version details
- Authenticates securely using an API key (environment variable)
- Returns structured JSON responses suitable for LLM tool integrations
- Easy to run locally for development or to integrate with desktop MCP clients
- Compatible with MCP clients such as Claude Desktop and the MCP Inspector
Prerequisites
- Node.js (recommended current LTS)
- Magic Meal Kits API key
- Optional: MCP client (e.g., Claude Desktop) for using the tool from an assistant
Installation / Configuration
Install via Smithery (recommended for Claude Desktop integration):
Optional global install via npm:
Development clone and run:
# Set MMK_API_KEY and MMK_API_BASE_URL in .env
Environment variables
- MMK_API_KEY — your Magic Meal Kits API key
- MMK_API_BASE_URL — base URL for the Magic Meal Kits API (example: https://magic-meal-kits.example.app)
Claude Desktop configuration example
Add to the “mcpServers” section of claude_desktop_config.json:
Current published version: 1.0.17
Available Tools
| Tool name | Description | Parameters |
|---|---|---|
| magic_meal_kits_server_version | Returns Magic Meal Kits API server version | None |
Example tool call (conceptual)
Request: “What version is the Magic Meal Kits server running?”
MCP response (JSON example):
Use Cases
- Assistant diagnostics: An assistant can check the backend version before explaining API behavior or troubleshooting user-reported issues.
- Deployment verification: Part of a deployment pipeline or post-deploy chatops flow can verify the running version matches the expected release.
- Conditional messaging: If the server reports an older version or incompatible build, the assistant can modify its guidance (e.g., flag missing features).
- Monitoring integration: Use the MCP server from an orchestration script to periodically confirm version and health as part of a lightweight monitoring check.
Concrete example
- A user asks the assistant why a particular endpoint behaves differently. The assistant calls magic_meal_kits_server_version, sees the API is on an older release, and returns: “Your backend is running v2023.11.2; the endpoint behavior you expect was introduced in v2024.01.0.”
Debugging & Troubleshooting
Run the server with console output for debugging:
# debug script provided by the project
# or if installed globally
Use the MCP Inspector to connect and interactively test tools:
# or
Common issue: “Server does not support tools (required for tools/call)”
- Ensure you have mmk-mcp version 1.0.11 or later. Update with:
Security note: Treat MMK_API_KEY as a secret. Do not commit it to source control, and prefer runtime environment injection or a secure secrets manager when deploying.
Additional Resources
- Repository: https://github.com/pureugong/mmk-mcp
- Smithery server listing: https://smithery.ai/server/@pureugong/mmk-mcp
- MCP Inspector: https://www.npmjs.com/package/@modelcontextprotocol/inspector
This MCP server is intentionally minimal: it focuses on providing reliable version metadata so AI assistants can adapt their behavior to the current backend state.