# The Met Museum MCP > The Met Museum MCP Server connects your AI client directly to over 470,000 artworks in one of the world's largest collections. You search by artist, filter by department, or narrow down pieces from specific centuries. It retrieves full metadata—including dimensions, materials, and open-access images—allowing deep historical and scholarly analysis without needing an API key. ## Overview - **Category:** knowledge-management - **Price:** Free - **Tags:** art-history, cultural-heritage, metadata-search, digital-archives, museum-collections, image-retrieval ## Description The Met Museum MCP Server connects your AI client directly to over 470,000 artworks in one of the world's deepest collections. You don't need an API key or fancy database setup; you just talk to it like you're talking to a curator. Your agent handles the heavy lifting. **General Search and Filtering** The `search_objects` tool runs across the entire Met collection, letting you narrow down millions of records using multiple filters at once. You can specify a department, define a date range, or filter by material type to quickly pinpoint what you need. If you know you need high-resolution images for your project, use the `search_with_images` tool; it guarantees that every object ID returned has available digital files attached. For broader scoping, start with `get_departments`, which gives you a full list of major museum sections—like Asian Art or Egyptian Art—so you can scope your search instantly. **Targeted Collection Scopes** You're not always looking at the whole lot. If you want to focus only on objects from one specific area, `get_objects_by_department` returns a comprehensive list of object IDs belonging solely to that department. Need to check out pieces created during a certain period? The `search_by_century` tool lets you find all artwork IDs within a defined century range—say, 1700 through 1799—cutting down decades of research in seconds. For the ultimate quick look at what's currently on display, use `search_on_view`, which locates object IDs for pieces actually hanging in the museum's physical galleries right now. **Deep Dive and Curation Tools** Sometimes you don't need a search; you need specifics. The `get_object` tool pulls all metadata—the title, artist, date, medium, dimensions, and images—for one unique artwork ID. It gives you the full scholarly package on any single piece. When you’re trying to nail down significance, use `search_highlights`. This retrieves object IDs for works that curators have specifically marked as important or noteworthy pieces across the whole collection. **Putting it Together: A Workflow Example** You can combine these functions easily. You might want to find all objects in Egyptian Art (`get_objects_by_department`), then filter those results down only to pieces created between 2500 BCE and 2400 BCE (`search_by_century`). Then, if you need the full details on the top five of those matches, you pass those IDs to `get_object`. It's a layered process that gives deep historical insight without writing a single SQL query. The system handles all the connections. The Met Museum MCP Server makes scholarly analysis simple. You talk naturally; your agent does the complex database work. Don't get bogged down in API calls or manual data fetching. Just tell it what you want, and it delivers the full metadata package for those 470,000 pieces. ## Tools ### get_departments Gets a list of all major museum departments, like Asian Art or Egyptian Art. ### get_object Gathers full metadata—title, artist, date, medium, dimensions—for one specific artwork ID. ### get_objects_by_department Returns a list of object IDs that belong to a single department. ### search_by_century Finds all artwork IDs created within a specific century range (e.g., 1500-1599). ### search_highlights Retrieves object IDs for works selected by curators as significant pieces. ### search_objects Runs a general search across the collection, allowing filtering by department, date range, or medium. ### search_on_view Finds object IDs for artworks that are currently displayed in the museum's galleries. ### search_with_images Searches for objects, ensuring they have available high-resolution images attached to their record. ## Prompt Examples **Prompt:** ``` Search for paintings by Monet. ``` **Response:** ``` Found 84 Monet paintings in the Met collection including: Water Lilies (1906), The Japanese Footbridge (1920), Rouen Cathedral series, and many Impressionist masterpieces with high-resolution images. ``` **Prompt:** ``` Show me the highlights from Egyptian Art. ``` **Response:** ``` Found 200+ Egyptian Art highlights including: Temple of Dendur (reconstructed in the museum), statues of pharaohs, sarcophagi, jewelry and papyrus scrolls spanning 3,000 years of Egyptian civilization. ``` **Prompt:** ``` Find sculptures from the 1800s. ``` **Response:** ``` Found 300+ sculptures from 1800-1899 including works by Rodin, Carpeaux and other 19th century masters. Each with dimensions, materials, credit lines and images. ``` ## Capabilities ### Search by Department Lists all major departments, allowing you to scope your search instantly. ### Retrieve Object Details Fetches complete metadata for a single artwork using its unique object ID. Includes images and dimensions. ### Filter by Department Scope Finds all available object IDs belonging to one specific department. ### Search by Century Range Filters the collection to find artworks created within a defined century (e.g., 1700-1799). ### Find Featured Works Retrieves a list of curator-selected 'highlight' objects from the entire collection. ### General Art Search Searches the full Met collection using multiple filters like department, date range, and medium type. ### Find On-View Items Locates objects that are currently displayed in the museum's physical galleries. ## Use Cases ### Tracking Renaissance Painting Trends A researcher needs to compare early Italian and French paintings. Instead of general search, they use `get_departments` to find both 'Italian Paintings' and 'French Paintings', then run `search_objects` on the resulting IDs. This immediately scopes the comparison to only relevant works. ### Curating a Time-Sensitive Exhibit An educator is building a lesson on American art from the 19th century. They use `search_by_century` to target objects between 1800 and 1899, then run `get_objects_by_department` (if they know the department) to limit results to 'American Art' only. ### Checking for Digital Assets A digital archivist needs a massive list of open-access images. They use `search_with_images` to filter the entire collection, ensuring every object they get back has a usable CC0 image file available for public domain use. ### Researching Current Museum Layouts A museum consultant needs to know which artifacts are physically on display this month. They run `search_on_view`. This gives them immediate insight into the current physical inventory, separate from the entire digital archive. ## Benefits - **Analyze Provenance:** Use `search_by_century` to filter thousands of artworks created in specific decades. This lets you track how artistic styles changed over time, making trend analysis easy. - **Pinpoint Specific Objects:** Once your search gives you a list of IDs, use `get_object` on the individual ID to get every detail: dimensions, medium, and credit line. No guessing required. - **Scope Your Search Effort:** Instead of searching 470k items generally, run `get_departments` first. Then, pipe that department into `get_objects_by_department` to narrow the focus immediately. - **Plan a Visit:** Use `search_on_view` if you need to know what's physically in the museum right now. This tool tells you what pieces are currently on display, so you don’t waste time searching for something that moved. - **Focus on Visual Results:** If your goal is visual output (e.g., a presentation slide), use `search_with_images`. It guarantees every returned object has an image attached, saving you from dead ends. ## How It Works The bottom line is that you talk to your agent using plain English, and the server handles all the complex filtering against 470k records for you. 1. Subscribe to this server on Vinkius. You won’t need an API key—all data is open access. 2. Your AI client (Claude, Cursor, etc.) runs the query against the available tools based on your natural language request. 3. The system returns a list of object IDs and their associated metadata, which you can then use to pull full details or perform follow-up searches. ## Frequently Asked Questions **Do I need an API key?** No! The Met Museum API is completely free and open. No authentication required. 80 requests/second rate limit. **Can I use the artwork images commercially?** Yes! Open Access images are licensed under CC0 (Public Domain). You can use them for any purpose without restriction. **How many artworks are in the collection?** The Met collection includes 470,000+ artworks spanning 5,000 years of art history from every part of the globe. **What departments are available?** The Met has 20+ departments including: European Paintings, Egyptian Art, Asian Art, Arms and Armor, Greek and Roman Art, Islamic Art, Modern and Contemporary Art, and many more. Use get_departments to see the full list. **When I use the get_object tool, what specific metadata fields do I receive?** It returns comprehensive details including title, artist, culture, date, medium, dimensions, and credit line. You also get image URLs for all Open Access CC0 public domain images. **If I want to find objects from a specific department, should I use the get_departments tool first?** Yes, you must run get_departments before calling get_objects_by_department. This initial step gives you the necessary department ID, which you then pass to retrieve object IDs. **Does search_by_century return full artwork details or just identifiers?** search_by_century only returns object IDs. To get the full metadata and images for those objects, you must follow up by using the get_object tool with those specific IDs. **What if I need to find pieces that have high-quality images, regardless of department?** Use search_with_images. This tool supports all standard filters (like date or medium) but specifically limits results to objects confirmed to have available image data.