Met Museum MCP

This MCP server plugs Claude directly into the Metropolitan Museum of Art's open-access collection—around 470,000 objects with high-resolution images and metadata. You get three tools: search for objects by keyword, pull detailed info on specific pieces, and list all departments. The API is public and needs no authentication, so setup is genuinely zero-friction. I'd reach for this when building educational tools, art reference databases, or anything that needs verifiable visual content. A history teacher could ask Claude to find Renaissance paintings of specific themes and get back object IDs, titles, artist names, dates, and direct image URLs. A designer sourcing period-accurate references can query by medium, culture, or time period. The metadata is rich—accession numbers, provenance, dimensions, credit lines—so it's not just pretty pictures. The quirks are mostly about the Met's API itself. Search results cap at 80 objects, so broad queries like "sculpture" will truncate. You can't filter by date range or medium in the initial search—you get everything matching your keyword, then filter manually or refine your query. Image URLs point to the Met's CDN, which is stable but not guaranteed forever. Some objects have no images even if they're marked open-access, because digitisation is ongoing. Performance is solid. Searches return in under a second, and the department list is instant. The object detail tool fetches full records including all available image assets, which is where the real value sits. If you're building a curriculum app or a visual research assistant, this is a clean, reliable data source. If you need comprehensive art history coverage beyond one institution, you'll want to pair it with other museum APIs or datasets. Don't bother if you need contemporary art (the Met skews historical), non-Western collections as a primary focus (though they exist), or if you're after auction estimates and market data. This is a research tool, not a commercial one.