Use WikiOO.org from your AI assistant

The server at a glance

Connect

Add this address to your assistant as a remote MCP server or custom connector:

https://direct.wikioo.org/mcp/

In an assistant with a settings screen, such as Claude or ChatGPT: open its connector settings, add a custom connector, paste the address and leave the authentication empty.

With Claude Code:

claude mcp add --transport http wikioo https://direct.wikioo.org/mcp/

In a client configured with a JSON file:

{
  "mcpServers": {
    "wikioo-art-search": {
      "type": "http",
      "url": "https://direct.wikioo.org/mcp/"
    }
  }
}

Ask for example

• Give me the biography of Artemisia Gentileschi with links to her works.
• Which artists were alive in Florence in 1500?
• What does the encyclopedia say about the origins of Cubism?

How a request runs

1. Your assistant searches the catalog and gets a list of results, each with an identifier (pk), a page address, an image address and a relevance score.
2. It reads the details of the result you care about: description, museum where the original is kept, and links to related pages.
3. For a price, it asks for a quotation at your size and in your currency, and gets the price with a link to an order form already filled in with that size and frame.
4. You open the link, check the order and pay on the website.

Try it from a terminal

Say hello to the server and read its name and instructions:

curl -s https://direct.wikioo.org/mcp/ \
  -H 'Content-Type: application/json' \
  -d '{"jsonrpc": "2.0", "id": 1, "method": "initialize"}'

List the tools with their full input schemas:

curl -s https://direct.wikioo.org/mcp/ \
  -H 'Content-Type: application/json' \
  -d '{"jsonrpc": "2.0", "id": 1, "method": "tools/list"}'

Call a tool:

curl -s https://direct.wikioo.org/mcp/ \
  -H 'Content-Type: application/json' \
  -d '{"jsonrpc": "2.0", "id": 1, "method": "tools/call", "params": {"name": "search_artworks", "arguments": {"query": "Artemisia Gentileschi", "limit": 3}}}'

Available tools

This list is built from what the server answers to tools/list, so it always matches the server. The descriptions are the ones your assistant reads, in English.

Which search tool to use:

• search_fts: You know a name or keywords and want the matching artworks, artists, museums and articles.
• search_artworks, search_artists, search_museums, search_articles: You want one kind of object only, with its own filters such as city, country or article category.
• search_sections_semantically: You have a question and want the passages of our texts that answer it.
• search_objects_semantically: You have a text and want the artworks, artists, museums or articles it is about.

All the tools:

• search_artworks: Search artworks
• search_museums: Search museums
• search_artists: Search artists
• search_articles: Search articles
• search_sections_semantically: Search passages by meaning (articles, museums, artists, artworks)
• search_objects_semantically: Find artworks, artists, museums or articles by meaning
• search_fts: Find artworks, artists, museums and articles by name or keywords
• get_detail_artwork: Get artwork details
• get_detail_artist: Get artist details
• get_detail_museum: Get museum details
• get_detail_article: Get article details
• get_a_quotation: Get a reproduction price quotation
• get_a_custom_reproduction_quotation: Get a custom reproduction price quotation (no catalog artwork needed)
• get_artwork_availability: Is this artwork available now, and at what price (stock, auction, originals, collectibles)

Good to know

• Search covers the whole catalog we share across our sites, not only the specialty of this site, so a result can fall outside it.
• Every address in an answer is absolute and points to this website.
• Each side of a reproduction must be 25 to 250 cm. A size outside that range is quoted at the nearest valid size, and the answer says what changed.
• A quotation is the price of a copy made to order. When the artwork itself is for sale now, the answer says so too, with its price and stock.
• A semantic search can answer index_unavailable: the assistant then falls back on the full-text search tools.

Errors

Privacy

The server asks for no name, no account and no key, and these tools read public catalog data only.

What we record. For each tool call we record the tool name, the language, the website, the time, the client's user agent and a salted hash of its IP address. Your search words are not part of this record. Our web server also keeps standard access logs, which contain IP addresses.

Why. To count usage of the server, and to keep it secure and working.

Who receives it. This data stays on the servers we rent from our hosting provider. We do not sell it and we do not share it with advertisers. Your assistant's provider handles your conversation under its own terms.

How long. One year at most, after which records are deleted automatically. Backup copies can last a few weeks longer.

Your choices. You can use the server without an account, and remove it from your assistant at any time. These records carry no name. To ask a question or request deletion, write to our support address below and tell us your IP address and the dates concerned.

Read the privacy policy of this website

Support

A question or a problem with the server? Write to us: [email protected]