wikipedia-mcp-server — Claude Code MCP server: setup and tools
Connects AI assistants to Wikipedia editions across languages through MCP. It fetches article summaries, targeted sections, citations, and nearby places.
Apache-2.0 6,422 installs ★ 6 updated 2026-10-24
claude mcp add --transport http wikipedia-mcp-server https://wikipedia-mcp-server--cyanheads.run.toolsWhat wikipedia-mcp-server does
The server exposes seven tools to query encyclopedia entries without scraping HTML pages. Models can run CirrusSearch queries with filters like incategory: or intitle:, request short summaries, and look up language editions. It also parses bibliographic references, giving direct access to cited URLs, DOIs, and archive links.
Articles arrive as clean text with intact heading tags. By default, pages larger than WIKIPEDIA_ARTICLE_OVERFLOW_BYTES (80,000 bytes) are cut back to an outline, requiring callers to fetch specific parts via section_index. The lead section always lives at index 0.
What it helps with
- search Wikipedia entries with query operators and spelling hints
- read introductory article summaries alongside Wikidata identifiers
- extract specific sections by index to stay within context budgets
- discover geotagged locations within a radius of given coordinates
- retrieve full bibliographic footnotes including DOIs and web links
Example prompts for the agent
- Look up the Wikipedia article on the Treaty of Westphalia and return its lead summary.
- Get the table of contents for quantum computing and fetch section index 3.
- Find geotagged Wikipedia pages within 5 kilometers of latitude 48.8566 and longitude 2.3522.
How to install wikipedia-mcp-server
Platform CLI: remote server
claude mcp add --transport http wikipedia-mcp-server https://wikipedia-mcp-server--cyanheads.run.toolsPlatform CLI: remote server
codex mcp add wikipedia-mcp-server --url https://wikipedia-mcp-server--cyanheads.run.toolsConfig file
{
"mcpServers": {
"wikipedia-mcp-server": {
"url": "https://wikipedia-mcp-server--cyanheads.run.tools"
}
}
}Config file
{
"servers": {
"wikipedia-mcp-server": {
"type": "http",
"url": "https://wikipedia-mcp-server--cyanheads.run.tools"
}
}
}Platform CLI: remote server
gemini mcp add -t http wikipedia-mcp-server https://wikipedia-mcp-server--cyanheads.run.toolsStep by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about wikipedia-mcp-server
How do I install wikipedia-mcp-server in Claude Code?
Run claude mcp add --transport http wikipedia-mcp-server https://wikipedia-mcp-server--cyanheads.run.tools. This connects the hosted endpoint directly to your workspace.
What capabilities does wikipedia-mcp-server provide?
It searches articles, reads full texts or selected chapters, and lists citations with external references. The server can also discover geotagged entries by coordinates and switch between language editions.
How does the server handle very long articles?
When an entry exceeds 80,000 bytes, the tool returns an outline instead of full prose. You can then request individual chapters by passing their section_index to wikipedia_get_article.