WikidataMCP

Overview

The Wikidata MCP server offers a comprehensive set of tools designed for AI agents and workflows to interact with Wikidata. It includes vector search capabilities for finding entities and properties, statement retrieval for detailed entity information, hierarchy exploration for understanding relationships, and SPARQL query execution for structured data retrieval. The server is well-documented with clear examples and usage guidelines.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• search_itemssearch_items(query: str, lang: str = "en") -> str

  Searches Wikidata items using vector search, returning QIDs with labels and descriptions.

  Try:Find information about the Eiffel TowerTry:Search for French impressionist paintersTry:Look up entries about quantum computers
• search_propertiessearch_properties(query: str, lang: str = "en") -> str

  Searches Wikidata properties using vector search, returning PIDs with labels and descriptions.

  Try:Find the property for 'capital of'Try:Search for properties related to populationTry:Look up properties about date of birth
• get_statementsget_statements(entity_id: str, include_external_ids: bool = False, lang: str = "en") -> str

  Returns direct statements for an entity in triplet-like text form.

  Try:Get all statements for entity Q535Try:Show statements for the Louvre MuseumTry:Retrieve statements for entity Q41298 with external IDs
• get_statement_valuesget_statement_values(entity_id: str, property_id: str, lang: str = "en") -> str

  Returns all statement values for an entity-property pair including qualifiers and references.

  Try:Get all values of 'population' for entity Q1860Try:Show statements with 'capital' property for FranceTry:Retrieve values of 'inception date' for entity Q535
• get_instance_and_subclass_hierarchyget_instance_and_subclass_hierarchy(entity_id: str, max_depth: int = 5, lang: str = "en") -> str

  Retrieves hierarchical context using 'instance of' and 'subclass of' relationships.

  Try:Show hierarchy for mammals (Q146)Try:Get hierarchy for programming languages (Q9143)Try:Display instance hierarchy for dogs with max depth 3
• execute_sparqlexecute_sparql(sparql: str, K: int = 10) -> str

  Executes a SPARQL query against Wikidata and returns up to K rows as CSV.

  Try:Find all capital cities in EuropeTry:Query for people born in Paris with SELECT * WHERETry:Get countries with population > 10 million

Comparable tools

Installation

uv run python main.py

docker compose up --build

{
  "mcpServers": {
    "wikidata": {
      "command": "curl",
      "args": ["https://wd-mcp.wmcloud.org/mcp"]
    }
  }
}

FAQ

What search capabilities does this MCP server provide?

The server offers both vector search and keyword search for finding Wikidata items (QIDs) and properties (PIDs), with fallback capabilities when vector search isn't available.

Can I execute custom SPARQL queries with this server?

Yes, the server includes an execute_sparql tool that allows running custom SPARQL queries against Wikidata, returning results as CSV text.

Compare WikidataMCP with