Wikipedia MCP Server
A Wikipedia MCP server: an agent searches articles and pulls full text, sections, links and summaries in the language you need
- 303 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Every request goes to the public Wikipedia API and the query text is visible to that service
- If you enable a network mode without auth, the server port is open to anyone who can reach it
Install
- Works in
Manual install
pipx install wikipedia-mcpThe recommended route for Claude Desktop: the wikipedia-mcp command becomes available in PATH globally.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Every request goes to the public Wikipedia API and the query text is visible to that service
- If you enable a network mode without auth, the server port is open to anyone who can reach it
This is third-party code. Review the repository files before installing.
What it does
Wikipedia MCP Server gives an agent Wikipedia content through 10 tools. search_wikipedia finds articles by query, and get_article returns the full text with sections, links and categories. get_sections, get_links and get_related_topics break down an article's structure and nearby topics. For precise answers there are three summary tools. summarize_article_for_query summarizes an article for a specific question, summarize_article_section takes one section, and extract_key_facts pulls out key facts. The language is set at launch, for example `wikipedia-mcp --language ja`, or through a country such as US, Japan or Germany. Language variants are supported: simplified and traditional Chinese, Serbian Latin and Cyrillic. The --enable-cache flag turns on a cache and lowers the load on the API. By default the server runs over stdio for Claude Desktop, and for networks there are http and streamable-http modes.
Who it is for. For people who ask an agent to check facts, find a definition or put together a brief from encyclopedia articles.
Good fit when
- You need to check a date, name or definition against Wikipedia before putting it into a text
- You want a short summary of an article for a specific question without reading the whole page
- You work with the Russian, Japanese or Chinese Wikipedia and want to pick the language at launch
- You need related topics and links inside an article to prepare reference material
- You need a place's coordinates from an article, which the get_coordinates tool returns
Not a fit when
- You need a search across the whole web and news: the server reads Wikipedia only
- You need data that Wikipedia lacks or that is outdated there
- You need to work offline: every request goes to the Wikipedia API, and the cache holds only answers already fetched
Example request
Find the article on Prince Shotoku in the Japanese Wikipedia, summarize the politics section and list related topicsMCP
- Transport
- stdio, http, sse
- Authentication
- not required
Environment variables
WIKIPEDIA_ACCESS_TOKENsecret- Wikipedia personal access token that lifts limits on frequent requests
Limitations
You need internet access and access to the Wikipedia API. Frequent requests get a 403 error from the API, and then you need a personal access token. The server passes article content as is and does not check its accuracy, so verify disputed facts against the article's own sources. All ten tools are read-only: you cannot edit articles through the server. In network mode, protect the port with a token or JWT if it is reachable from outside.
How to disable. Remove the wikipedia entry from the MCP client config and, if you like, run `pipx uninstall wikipedia-mcp`. If you set a personal Wikipedia token, you can revoke it in your Wikipedia account settings.
FAQ
How do I connect Wikipedia to Claude Desktop?
Install the server with `pipx install wikipedia-mcp` and add a wikipedia entry with the command wikipedia-mcp to claude_desktop_config.json. If Claude cannot see the command, give the full path that `which wikipedia-mcp` prints.
How do I switch the server to the Russian Wikipedia?
Start it with `--language ru` or `--country RU`. The `wikipedia-mcp --list-countries` command prints the supported countries.
What do I do about 403 errors when querying Wikipedia?
Create a Wikipedia personal access token and pass it with the --access-token flag or the WIKIPEDIA_ACCESS_TOKEN variable. With a token the rate limit is looser.
How do I protect a network server with a password?
Start it with `--transport http` and add `--auth-mode static --auth-token <token>`. The other option is `--auth-mode jwt` with --auth-jwks-uri and --auth-issuer.
How does summarize_article_for_query differ from get_summary?
get_summary returns a general short summary of the article. summarize_article_for_query takes a question and summarizes the article for it, with the length set by max_length, 250 by default.
Discussion
No questions or reviews yet.
Related
MCP serversA web search MCP server that gives the agent full page text at once: Stack Overflow answers, GitHub threads, arXiv papers
An MCP server that searches Google and Scholar, reads pages and PDFs and stores findings in a local project knowledge graph
A PRAW-based Reddit MCP server: an agent searches posts, reads user and subreddit history and can post as you if allowed
MCP server for Reddit: an agent reads hot threads from a subreddit and analyzes posts with their comments
Web search via DuckDuckGo and page reading for the agent with no API keys, with built-in rate limiting
A Reddit MCP server: the agent browses posts, searches content and analyzes user activity with no fluff
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Wikipedia MCP Server (this entry) | no | not required | 303 | |
| Brave Search MCP Server | yes | API key | 1,544 | searches the whole web, news and images; Wikipedia MCP reads only Wikipedia articles |
| CoexistAI | no | not required | 530 | a local research server with web search, Reddit, YouTube and maps; here there is one source, Wikipedia |
README in short
The README starts with install options through pipx, PyPI, Smithery and Docker and a Claude Desktop config example. Parameters for each tool follow, then a table of countries and language variants.