MediaWiki API MCP Server
Enables LLMs to interact with MediaWiki installations as a bot user, supporting page editing, searching, moving, deleting, comparing revisions, and retrieving site information through the MediaWiki API.
README
MediaWiki API MCP Server
A Model Context Protocol (MCP) server that allows LLMs to interact with MediaWiki installations through the MediaWiki API as a bot user.
Features
Tools
The server provides various MCP tools with the wiki_ prefix:
| Tool | Description |
|---|---|
wiki_page_edit |
Edit or create MediaWiki pages with comprehensive editing options |
wiki_page_get |
Retrieve page information and content |
wiki_page_parse |
Parse page content with support for wikitext processing, HTML generation, metadata extraction, and advanced parsing features |
wiki_page_compare |
Compare two pages, revisions, or text content to show differences between them |
wiki_page_move |
Move pages with support for talk pages, subpages, and redirects |
wiki_page_delete |
Delete pages with support for talk pages, watchlist management, and logging |
wiki_page_undelete |
Undelete (restore) deleted MediaWiki pages with comprehensive restoration options |
wiki_search |
Search for pages using MediaWiki's search API with advanced filtering |
wiki_opensearch |
Search using OpenSearch protocol for quick suggestions and autocomplete |
wiki_meta_siteinfo |
Get overall site information including general info, namespaces, statistics, extensions, and more |
Installation
-
Clone the repository and checkout the
mainbranch -
Install dependencies using UV:
uv install
- Set up environment variables:
export MEDIAWIKI_API_URL="http://mediawiki.test/api.php"
export MEDIAWIKI_API_BOT_USERNAME="YourUserName@YourBotName"
export MEDIAWIKI_API_BOT_PASSWORD="YourBotPassword"
export MEDIAWIKI_API_BOT_USER_AGENT="MediaWiki-MCP-Bot/1.0 (your.email@mediawiki.test)" # Optional
Usage
Running the Server
uv run mediawiki-api-mcp
Or directly:
python -m mediawiki_api_mcp.server
Configuration with Claude Desktop
Configuration File Location
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
Template Configuration
Add to your Claude Desktop configuration file:
{
"mcpServers": {
"mediawiki-api": {
"command": "uv",
"args": [
"--directory",
"/absolute/path/to/mediawiki-api-mcp",
"run",
"mediawiki-api-mcp"
],
"env": {
"MEDIAWIKI_API_URL": "http://mediawiki.test/api.php",
"MEDIAWIKI_API_BOT_USERNAME": "YourUserName@YourBotName",
"MEDIAWIKI_API_BOT_PASSWORD": "YourBotPassword",
"MEDIAWIKI_API_BOT_USER_AGENT": "MediaWiki-MCP-Bot/1.0 (your.email@mediawiki.test)"
}
}
}
}
Configuration Instructions
- Replace
/absolute/path/to/mediawiki-api-mcpwith the actual absolute path to this project directory - Update
MEDIAWIKI_API_URLwith your MediaWiki installation's API endpoint - Set
MEDIAWIKI_API_BOT_USERNAMEto your bot username (typically in formatYourUserName@YourBotName) - Set
MEDIAWIKI_API_BOT_PASSWORDto the generated bot password from your wiki'sSpecial:BotPasswordspage - Customize
MEDIAWIKI_API_BOT_USER_AGENTwith appropriate contact information (optional)
Bot Password Setup
Create bot credentials at e.g.: http://mediawiki.test/index.php/Special:BotPasswords
Required permissions:
- Basic rights: Read pages
- High-volume editing: Edit existing pages, Create, edit, and move pages
- Additional permissions as needed for your specific use case
Security Notes
- Keep your bot credentials secure and never commit them to version control
- Use the principle of least privilege when setting bot permissions
- Monitor bot activity through your MediaWiki's logging interface
- Consider using IP restrictions for additional security
Development
Technology Stack
- MCP SDK: mcp >= 1.2.0 (using FastMCP pattern)
- HTTP Client: httpx >= 0.27.0 for MediaWiki API calls
- Data Validation: pydantic >= 2.0.0 for configuration models
- Environment: python-dotenv >= 1.0.0 for configuration
- Testing: pytest with pytest-asyncio for async testing
- Code Quality: ruff for linting, mypy for type checking
Architecture
- FastMCP server with
@mcp.tool()decorators - Separation of concerns: server → handler → client → MediaWiki API
- AsyncIO throughout for non-blocking operations
- Environment-based configuration for MediaWiki credentials
Client Layer (client.py and client_modules/)
- Handles MediaWiki API authentication and requests
- Manages CSRF tokens and session state
- Provides typed methods for API operations
Tools Layer (tools/)
- Defines MCP tool schemas using JSON Schema
- Separated by functional area (edit, search)
- Ensures all tools have
wiki_prefix
Handlers Layer (handlers/)
- Implements actual tool logic
- Handles argument validation and error handling
- Returns properly formatted MCP responses
Server Layer (server.py and server_tools/)
- Main MCP server orchestration
- Routes tool calls to appropriate handlers
- Manages configuration and client lifecycle
Project Structure
The project is organized into modular components for maintainability:
mediawiki-api-mcp/
├── mediawiki_api_mcp/
│ ├── __init__.py
│ ├── server.py # FastMCP server with tool definitions
│ ├── client.py # MediaWiki API client
│ ├── client_modules/ # Client modules for API operations
│ │ ├── __init__.py # Client module exports
│ │ └── client_*.py # Individual client modules
│ ├── handlers/ # Business logic handlers
│ │ ├── __init__.py # Handler exports
│ │ └── wiki_*.py # Individual tool handlers
│ ├── tools/ # Tool schemas
│ │ ├── __init__.py # Tool schema exports
│ │ └── wiki_*.py # Individual tool schemas
│ └── server_tools/ # Tool definitions
│ ├── __init__.py # Tool definition exports
│ └── wiki_*.py # Individual tool definitions
├── tests/ # Test suite
│ └── test_*.py # Test files matching handlers
├── docs/ # Documentation
│ └── tools/ # Tool-specific documentation
│ └── wiki_*.md # Individual tool docs
├── pyproject.toml # Project configuration
└── uv.lock # Dependency lock file
Running Tests
uv run pytest
Run specific test modules, e.g.:
uv run pytest tests/test_server.py
uv run pytest tests/test_tools.py
uv run pytest tests/test_wiki_page_edit.py
uv run pytest tests/test_wiki_search.py
uv run pytest tests/test_wiki_opensearch.py
Code Quality
# Linting
uv run ruff check
# Type checking
uv run mypy mediawiki_api_mcp
Error Handling
The server implements comprehensive error handling:
- Configuration errors: Missing environment variables
- Authentication errors: Invalid credentials or permissions
- API errors: Network issues, invalid requests
- Tool errors: Missing parameters, invalid arguments
All errors are returned as MCP TextContent responses for LLM visibility.
Security
- Bot credentials are required for editing operations
- All API requests include proper User-Agent headers
- CSRF tokens are automatically managed
- Input validation on all tool parameters
License
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。