MCP.so
ログイン
サーバー

Met Museum MCP Server

@mikechao

Met Museum MCP integration to discover the art collection at The Metropolitan Museum of Art in New York

概要

What is Met Museum MCP Server?

A Model Context Protocol (MCP) server that provides AI models access to the Metropolitan Museum of Art Collection through natural language interactions. It enables searching The Met's art collection and retrieving artwork details, including images, via tool results.

How to use Met Museum MCP Server?

Run directly with npx -y metmuseum-mcp to start the server using stdio transport (default for desktop clients). For ChatGPT, use the --http flag and expose the server via a tunnel like ngrok. Client-specific setup instructions are provided for Claude Desktop, ChatGPT, and LibreChat.

Key features of Met Museum MCP Server

  • List all valid departments at The Met
  • Search museum objects with filters (query, images, title, department, pagination)
  • Retrieve full object details including Open Access images (base64 JPEG)
  • Launch an interactive Met Explorer app from MCP clients
  • Supports both stdio and Streamable HTTP transports

Use cases of Met Museum MCP Server

  • Explore works of a specific artist like Vincent van Gogh
  • Find paintings from a specific department (e.g., Asian Art)
  • Look up detailed information about a specific artwork by title or ID
  • Browse the Met collection interactively through an in-client app
  • Search for objects matching keywords like "cat" or "sunflowers"

FAQ from Met Museum MCP Server

What data does it access?

It uses the Metropolitan Museum of Art Collection API to retrieve department lists, search results, and full object records including images under Open Access. It is a third-party implementation, not officially associated with The Met.

What are the runtime requirements?

Node.js v18 or later is required. The server can run directly via npx with no local installation needed.

What transports are available?

Two transports are supported: stdio (default, for desktop clients like Claude Desktop) and Streamable HTTP (enable with --http, for ChatGPT and other web clients).

How are images returned?

When retrieving an object and returnImage is true (default), the Open Access image is included in the tool result as a base64-encoded JPEG. Otherwise, the primary image URL is provided.

Is authentication required?

No authentication is mentioned for the MCP server itself. When using with ChatGPT, "No Auth" is selected as the authentication type for the connector.

タグ

「その他」の他のコンテンツ