home / mcp / harvard art museums mcp server
Provides access to Harvard Art Museums data via MCP, enabling search and retrieval of artworks, objects, and people.
Configuration
View docs{
"mcpServers": {
"alexlin1234-harvard-art-museums-mcp-server": {
"command": "python",
"args": [
"/absolute/path/to/harvard-art-museums/server.py"
],
"env": {
"HARVARD_ART_MUSEUMS_API_KEY": "your_api_key_here"
}
}
}
}The Harvard Art Museums MCP Server provides programmatic access to the Harvard Art Museums API, enabling you to search and retrieve information about artworks, artists, and museum objects through a lightweight, model-context protocol interface.
Connect to the Harvard Art Museums MCP Server from an MCP-compatible client and start querying the collection. You can search for artworks, fetch detailed object information, look up artists, retrieve person details, or obtain a random artwork, all within a single, coherent API layer.
Common workflows include: searching for artworks with filters (such as artist, period, or medium), getting the full data for a specific object by its ID, listing artists or people in the collection, and exploring a random piece with optional cultural or classification filters.
Prerequisites: you need Python installed and a working internet connection to access the Harvard Art Museums API.
1) Create and activate a virtual environment.
2) Install required Python packages.
3) Obtain your free API key from Harvard Art Museums and store it securely as an environment variable.
4) Prepare a configuration file or environment with your API key to run the MCP server.
5) Start the MCP server using the documented command and verify it is reachable from your MCP client.
The MCP server exposes five tools for interacting with the Harvard Art Museums data: search_objects, get_object, search_people, get_person, and random_object. Each tool corresponds to a specific operation in the API, allowing you to filter, retrieve, and explore data efficiently.
Security: protect your API key and limit access to trusted MCP clients. Do not expose the key in public code or shared configurations.
Usage tips: cache results when possible to respect rate limits (the API allows up to 2,500 requests per day per key). Use precise filters to minimize unnecessary requests and improve response times.
Search artworks in the collection with filters such as artist, period, or medium to quickly locate relevant items.
Fetch detailed information about a specific artwork by its object ID, including title, artist, date, and provenance.
Search for artists and people associated with the collection using name or role filters.
Fetch detailed information about a specific person by their ID, including biography and related works.
Return a randomly selected artwork, with optional filters for classification or culture to guide the result.