# met-museum-mcp-server [Health: Active]

**Category:** 🎨 Art & Culture  
**Repository:** https://github.com/cyanheads/met-museum-mcp-server  
**GitHub Stars:** 4  
**npm Downloads (last month):** 1045  
**Views:** 2  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/met-museum-mcp-server

## Description
MET Museum collection via MCP — 500K+ artworks, metadata, provenance, open-access images.

## Tools
Capabilities this server exposes over MCP:

- **met_list_departments** — Return all 19 curatorial departments with their numeric IDs and display names
- **met_search_collections** — Search the collection by keyword, optionally matched in titles or tags only, with filters for department, date range, medium, geography, image availability, on-view status, and highlight designation
- **met_list_objects** — List every object ID in a department, every object created or revised since a date, or both, without a keyword
- **met_get_object** — Fetch full records for one or more object IDs — metadata, provenance, artist info, CC0 image URLs, tags, and Wikidata links — optionally with up to 3 CC0 images as image content

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "met-museum-mcp-server": {
    "command": "npx",
    "args": ["-y","@cyanheads/met-museum-mcp-server"]
  }
}
```

## Documentation

## What met-museum-mcp-server does

The met-museum-mcp-server MCP server exposes the Metropolitan Museum of Art’s public collection through four MCP tools. It covers more than 500,000 works spanning roughly 5,000 years and organizes discovery around the Met’s 19 curatorial departments.

The server supports keyword searches, filter-based collection browsing, object-ID enumeration, and detailed record retrieval. Returned artwork data can include titles, artist information, dates, media, geography, provenance, public-domain indicators, CC0 image URLs, tags, and links to related Wikidata entities. Getty ULAN and AAT URLs may also be present in linked data.

This fits the Art & Culture category and is useful when an agent needs structured Met collection data rather than general web search results.

## How it works

`met_list_departments` returns department names and numeric IDs. Those IDs can then be supplied to search and listing operations.

`met_search_collections` requires a keyword, although `"*"` can be used for a filter-only search. It can narrow results by title or tags, department, medium, date range, geography, image availability, on-view status, and highlight status. Results support pagination through offsets and report totals, truncation, and the resolved query.

`met_list_objects` enumerates object IDs by department, by records created or revised since a date, or by both criteria. Without filters, it can enumerate the full collection. Lists are cached for up to one hour.

`met_get_object` retrieves records for object IDs returned by the search or listing tools. It accepts batches of up to 500 IDs but fetches up to 20 distinct records per call, reporting later IDs for follow-up. With `includeImages: true`, it can attach CC0 images for up to three eligible records as image content.

## Setup and configuration

A public Streamable HTTP instance is available at `https://met-museum.caseyjhand.com/mcp`, so a client can connect without installing the package. Local clients can run the package through a standard STDIO configuration using `bunx @cyanheads/met-museum-mcp-server@latest`. The README also provides installation links for Claude Desktop, Cursor, and VS Code.

The server supports STDIO and Streamable HTTP transports. Its underlying framework also documents authentication choices including none, JWT, and OAuth, plus several storage backends, but the provided client setup does not require additional environment variables.

## Tools and capabilities

The met-museum-mcp-server MCP server provides:

- Department discovery with valid numeric IDs
- Search pagination of up to 500 IDs per page through the first 10,000 matches
- Unkeyworded enumeration by department or update date
- Batch artwork retrieval with partial-success reporting
- Provenance and public-domain fields for image decisions
- CC0 image URLs and optional image content
- Typed reporting for rate limits, blocked access, unavailable upstream services, invalid filters, and other request failures
- Deferred-record reporting when structured output exceeds the stated byte budget

## Limitations and notes

Image fields are empty for objects that are not public domain or do not have CC0 images. The Met’s index can contain IDs that are no longer served, so individual retrievals may fail. Batch retrieval reports failed IDs separately when some records succeed.

Search results are limited to the first 10,000 matches, while object listings have no stated depth limit. The Cloisters and Medieval Art departments share a result set at the Met, so a page may contain fewer records than requested and include a notice. Recent updates may be absent from object listings for up to an hour because of caching. Upstream blocking, rate limiting, and service failures are surfaced as typed errors rather than raw Met error pages.

_Full upstream README: https://allmcps.com/mcp/met-museum-mcp-server/readme_