What it does

This server exposes Wikipedia search and article retrieval as MCP tools. It provides three endpoints: search articles by keyword in any language with configurable result limits, fetch full article details with optional wikitext markup, and retrieve random articles for exploration. For example, you can search for "machine learning" in English, fetch the full article, and pass it to Claude for analysis or citation.

Who it's for

Developers integrating Claude into research workflows or building fact-checking tools. Also useful for teams developing conversational agents that need to ground responses in verifiable encyclopedic sources and citations.

Common use cases

• Search Wikipedia for topic background during a conversation
• Retrieve article content for analysis or citation in LLM responses
• Discover related articles through random retrieval
• Build research tools that support multiple Wikipedia language editions
• Verify AI-generated claims with encyclopedic sources

Setup pitfalls

• Default language is Japanese (ja), not English — explicitly set lang: en for English Wikipedia
• Article content is returned as raw wikitext markup, not plain text — clients must parse or strip wiki syntax
• Wikipedia API enforces rate limits — implement backoff strategies if making many requests in production
• No CI tests and last commit was 350 days ago — monitor for compatibility if Wikipedia API changes