Integrations
MediaWiki MCP Server
The MediaWiki MCP server provides a powerful set of tools for AI assistants like Claude to interact with your MediaWiki instance. It uses nodemw to communicate with the MediaWiki API to retrieve MediaWiki page content.
Key Features
- Get page content : Get the HTML content and metadata of a wiki page
- Wiki Search : Search for pages based on keywords
- Get category information : Get the category information of a page
- Link Analysis : Extracting internal links within a page
Prerequisites
- Node.js 18 or higher
- npm or yarn
How to use
install
Example of use in Claude settings
Example configuration in the Claude config file: claude_desktop_config.json
Copy
Available Features
List of environment variables
Environment variable name | Required | Default value | explanation |
---|---|---|---|
MEDIAWIKI_PROTOCOL | × | https | Communication protocol with the MediaWiki server |
MEDIAWIKI_SERVER | ○ | - | The hostname of your MediaWiki server (e.g. ja.wikipedia.org) |
MEDIAWIKI_PATH | × | /w | MediaWiki paths |
MEDIAWIKI_USERNAME | × | - | Username for logging in |
MEDIAWIKI_PASSWORD | × | - | Login password |
MEDIAWIKI_USER_AGENT | × | MediaWiki-MCP-Server/1.0 | User agent when making API requests |
MEDIAWIKI_CONCURRENCY | × | 3 | Limiting the number of simultaneous connections |
DEBUG | × | 0 | Debug mode (enabled at 1) |
Known Issues
- If the article is too large, it may fail to retrieve the page (requires chunking).
- Depends on nodemw, so functionality is limited
link
This server cannot be installed
A tool that enables AI assistants like Claude to interact with MediaWiki instances by retrieving page content, performing searches, and analyzing wiki information through the MediaWiki API.