Wikidata Mcp Server
Pull structured facts, SPARQL query results, and external ID mappings from Wikidata inside agent-driven research sessions.
Overview
io.github.cyanheads/wikidata-mcp-server is a MCP server for the Idea phase that searches Wikidata, runs SPARQL, and resolves external identifiers for agent-led research.
What is this MCP server?
- Search and fetch Wikidata entities with agent-callable tools
- Execute SPARQL queries with WIKIDATA_SPARQL_TIMEOUT_MS default 55000
- Resolve external identifiers to Wikidata items for cross-source joins
- Stdio npm package plus optional streamable-http remote at wikidata.caseyjhand.com/mcp
- Wikimedia-compliant WIKIDATA_USER_AGENT default string
- Server version 0.1.9
- Default WIKIDATA_SPARQL_TIMEOUT_MS 55000
- Documented remote streamable-http endpoint https://wikidata.caseyjhand.com/mcp
Community signal: 1 GitHub stars.
What problem does it solve?
Unstructured web summaries mix fact and fiction when you need comparable entities, dates, and IDs for a new product thesis.
Who is it for?
Builders researching markets, mapping competitors to knowledge-graph IDs, or prototyping data-heavy content products.
Skip if: Real-time news monitoring, proprietary CRM enrichment, or workflows that cannot tolerate SPARQL latency and Wikimedia rate norms.
What do I get? / Deliverables
Your agent returns Wikidata-backed entities and query tables you can paste into PRDs, pitch outlines, and validation spreadsheets.
- Wikidata entity payloads with statements and labels
- SPARQL result sets for lists and filters
- External identifier mappings linked to Q-items
Recommended MCP Servers
Journey fit
How it compares
Structured knowledge-graph MCP access, not a general web search skill or Wikipedia prose reader.
Common Questions / FAQ
Who is io.github.cyanheads/wikidata-mcp-server for?
Solo builders and small teams who want MCP agents to query Wikidata entities and SPARQL during idea and validate research.
When should I use io.github.cyanheads/wikidata-mcp-server?
Use it when you need authoritative entity records, identifier crosswalks, or custom SPARQL result sets while scoping a product or content angle.
How do I add io.github.cyanheads/wikidata-mcp-server to my agent?
Add stdio config for @cyanheads/wikidata-mcp-server with Bun run start:stdio, set WIKIDATA_USER_AGENT per policy, or point compatible clients at the streamable-http remote URL when available.