mcp-server-wikipedia
This server provides a progressive retrieval strategy for Wikipedia, minimizing token usage by allowing LLMs to search, summarize, and selectively fetch sections. The tools include:
search_articles(query, limit=5): Search Wikipedia and return top matching pages with snippets.
get_summaries(titles): Fetch compact summaries for one or more page titles.
get_toc(title): Retrieve a structured table of contents with section indices, titles, and anchors.
get_section(title, section): Fetch a specific section by index or exact title.
get_page(title): Retrieve the full plain-text content of a page.
These tools can be chained (search → summaries → toc → section) to reduce token usage by up to 80% compared to fetching entire pages. It integrates as an MCP server with clients like Claude Desktop.
Provides tools for searching Wikipedia articles, getting summaries, inspecting table of contents, and retrieving specific sections or full pages using a progressive retrieval strategy to minimize token usage.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@mcp-server-wikipedia Search for light reactions and summarize top result"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Wikipedia MCP Server 📚
Surgical, token-efficient Wikipedia knowledge retrieval MCP server for AI agents with deterministic summary, search, and full-article projection.
🌐 Live Documentation & Web Portal: https://wikipedia.builditwithai.xyz
⚡ Quickstart
# 1-Line Universal Installer (Auto-configures Claude Desktop, Cursor, Claude Code, Antigravity, VS Code, Zed, Windsurf)
curl -fsSL "https://wikipedia.builditwithai.xyz/install" | bash
# Or run directly via your preferred runtime:
uvx mcp-server-wikipedia
npx -y mcp-server-wikipediaRelated MCP server: Wikipedia MCP Server
🤖 Client Setup
A. Claude Code (CLI)
claude mcp add wikipedia -- uvx mcp-server-wikipediaB. Cursor & Google Antigravity (mcp.json)
{
"mcpServers": {
"wikipedia": {
"command": "uvx",
"args": ["mcp-server-wikipedia"]
}
}
}C. Claude Desktop (claude_desktop_config.json)
{
"mcpServers": {
"wikipedia": {
"command": "uvx",
"args": ["mcp-server-wikipedia"]
}
}
}D. VS Code (Cline / Roo Code / Continue)
{
"mcpServers": {
"wikipedia": {
"command": "npx",
"args": ["-y", "mcp-server-wikipedia"]
}
}
}🛠️ Tools & Capabilities
Tool Name | Parameters | Description | Return Type |
|
| Fetches concise, factual lead summary of any Wikipedia page. |
|
|
| Fetches complete article body formatted in clean, structured Markdown. |
|
|
| Searches Wikipedia articles with autocomplete and relevance ranking. |
|
|
| Retrieves a specific section by header name, saving context window space. |
|
|
| Extracts outgoing cross-reference links from an article. |
|
|
| Lists taxonomic categories and classifications for a page. |
|
|
| Dynamically loads research playbooks from GitHub. |
|
| (none) | Lists all available Wikipedia research skills. |
|
🔒 Telemetry & Privacy
This package collects anonymous, non-PII diagnostic telemetry (command executions, latency, error codes) to improve tool reliability. No article search terms, personal data, source code, or environment variables are ever collected or stored.
You can opt out anytime by setting either of the following environment variables:
export DO_NOT_TRACK=1
# or
export MCP_TELEMETRY_OPT_OUT=1📄 License
MIT License. See LICENSE for details.
Available Tools
7 toolsget_pageBRead-onlyIdempotent
Fetch a full Wikipedia page as plain text.
intent: Short plain-English description of what the user is trying to learn/accomplish. E.g. "background on the Suez crisis for an essay", "verify a claimed date".
Inputs: title (required), intent. Returns: result
get_sectionBRead-onlyIdempotent
Fetch one section by section index or exact section title.
Inputs: title (required), section (required). Returns: result
get_summariesARead-onlyIdempotent
Fetch compact summaries for one or more Wikipedia page titles.
Inputs: titles (required). Returns: result
get_tocARead-onlyIdempotent
Return a page's table of contents as section index, title, and anchor.
Inputs: title (required). Returns: result
search_articlesARead-onlyIdempotent
Search English Wikipedia and return a compact list of candidate pages.
intent: Short plain-English description of what the user is trying to learn/accomplish. E.g. "background on the Suez crisis for an essay", "verify a claimed date".
Inputs: query (required), limit, intent. Returns: error, results
skill_readARead-onlyIdempotent
Read a usage skill (markdown) by name — see skills_list for names. Read 'interpreting-errors' after any error or empty result to recover.
Inputs: name (required). Returns: result
skills_listARead-onlyIdempotent
List usage skills for this server. Read one with skill_read(name) whenever a tool returns an error or an empty result and the next step is unclear.
Inputs: no named parameters. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 7 tools
The tools are mostly distinct: get_page retrieves full text, get_section retrieves a specific section, get_summaries retrieves summaries, get_toc returns the table of contents, and search_articles finds candidates. However, get_page and get_summaries could overlap slightly in use cases (full text vs. summary) and the skills_list/skill_read pair is clearly distinct. Overall, minimal overlap.
All tools consistently follow a verb_noun pattern with get_ prefix and lowercase nouns: get_page, get_section, get_summaries, get_toc, search_articles, skills_list, skill_read. The pattern is predictable and consistent.
7 tools is within the ideal 3-15 range. Each tool serves a clear purpose for interacting with Wikipedia: search, read full page, read section, get summary, get TOC, and skill management. No redundant tools.
The set covers the core Wikipedia read operations: search, full page, section, summary, TOC. Missing might be a way to get page metadata or links, but for a read-only reference server this is well-covered. No obvious dead ends.
Maintenance
Related MCP Connectors
LLM-ready web search + instant answers + URL-to-clean-text fetch for agents and RAG.
Web knowledge fetching tools for your agent.
Web search, fetch, extract, and research for AI agents. Markdown output + AI-synthesized answers.
Live Google web search for AI agents that returns only the passages that answer the question.
Related MCP Servers
- AlicenseAqualityCmaintenanceEnables LLMs to search for keywords and fetch full page content from Wikipedia across various languages. It provides direct access to Wikipedia information through search and fetch tools.24MIT
- AlicenseAqualityDmaintenanceEnables AI assistants to access Wikipedia content, search articles, retrieve historical events, and fetch images through the Wikipedia API.418 npm1MIT
- FlicenseNot gradedqualityDmaintenanceEnables searching and retrieving Wikipedia content across multiple languages, including article summaries, sections, coordinates, and related topics.-
- FlicenseNot gradedqualityDmaintenanceEnables searching and reading Wikipedia articles through tools like search, section listing, and content retrieval. Supports prompt templates and resource suggestions for topic exploration.-