ZIM MCP Server
Enables large language models to directly access and search content in ZIM files, allowing offline question answering and information retrieval from resources like Wikipedia.
README
ZIM MCP Server
ZIM (Zeno IMproved) is a file format developed by the nonprofit organization Kiwix, designed specifically for offline storage and access to Wikipedia and other large reference content. The ZIM format supports high compression rates and fast searching, enabling entire Wikipedia content to be compressed into relatively small files for convenient storage and use, especially in environments without internet connectivity.
ZIM MCP Server provides large language models with the ability to directly access and search content within ZIM files, allowing people to use local AI models for question answering and information retrieval from these offline knowledge resources, even without network access.
About Kiwix
Kiwix is a nonprofit organization dedicated to making online knowledge content (especially Wikipedia, TED talks, etc.) accessible offline. Kiwix has developed tools for creating, viewing, and searching ZIM files, through which people can package large amounts of online knowledge resources into ZIM files for local access. The Kiwix project is particularly important for developing countries and regions without internet connectivity, as it enables people in these areas to access rich knowledge resources, promoting the dissemination of knowledge and equal educational opportunities.
Installation
1. Clone the Repository
git clone https://github.com/ThinkInAI-Hackathon/zim-mcp-server.git
2. Install uv
- On Windows:
- If you have not set execution policy before, run:
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex" - Otherwise, you can run:
irm https://astral.sh/uv/install.ps1 | iex
- If you have not set execution policy before, run:
- On MacOS:
- Please refer to the official uv installation guide.
3. Install Dependencies
cd path\to\zim-mcp-server # (e.g., D:\zim-mcp-server)
uv sync
4. Prepare ZIM Files
Download ZIM files (e.g., Wikipedia, Wiktionary, etc.) from the Kiwix Library and place them in a directory, e.g.,
path\to\zim\files # (e.g., D:\ZIM)
Configuration for Claude Desktop
The configuration for Claude Desktop requires the local path to your cloned repository.
Add the following to your claude_desktop_config.json (replace the paths with your actual local paths):
"zim-mcp-server": {
"command": "uv",
"args": [
"--directory",
"path\\to\\zim-mcp-server", // e.g., D:\\zim-mcp-server
"run",
"server.py",
"path\\to\\zim\\files" // e.g., D:\\ZIM
]
}
"path\\to\\zim-mcp-server": Local path to your cloned zim-mcp-server repository (e.g., D:\zim-mcp-server)."path\\to\\zim\\files": Directory containing your ZIM files (e.g., D:\ZIM).
Claude Desktop will recognize the server only after this configuration.
Available Tools
list_zim_files - List all ZIM files in allowed directories
No parameters required.
search_zim_file - Search within ZIM file content
Required parameters:
zimFilePath(string): Path to the ZIM filequery(string): Search query term
Optional parameters:
limit(integer, default: 10): Maximum number of results to returnoffset(integer, default: 0): Starting offset for results (for pagination)
get_zim_entry - Get detailed content of a specific entry in a ZIM file
Required parameters:
zimFilePath(string): Path to the ZIM fileentryPath(string): Entry path, e.g., 'A/Some_Article'
Optional parameters:
maxContentLength(integer, default: 100000): Maximum length of returned content
Examples
Listing ZIM files:
{
"name": "list_zim_files"
}
Response:
{
"Found 2 ZIM files in 1 directories:
[
{
"name": "wikipedia_en_all_nopic_2023-07.zim",
"path": "D:/ZIM/wikipedia_en_all_nopic_2023-07.zim",
"directory": "D:/ZIM",
"size": "95123.45 MB",
"modified": "2023-08-01T12:00:00"
},
{
"name": "wiktionary_en_all_nopic_2023-07.zim",
"path": "D:/ZIM/wiktionary_en_all_nopic_2023-07.zim",
"directory": "D:/ZIM",
"size": "1234.56 MB",
"modified": "2023-08-01T12:30:00"
}
]
}
Searching ZIM files:
{
"name": "search_zim_file",
"arguments": {
"zimFilePath": "D:/ZIM/wikipedia_en_all_nopic_2023-07.zim",
"query": "artificial intelligence",
"limit": 3
}
}
Response:
Found 120 matches for "artificial intelligence", showing 1-3:
## 1. Artificial intelligence
Path: A/Artificial_intelligence
Snippet: Artificial intelligence (AI) is intelligence demonstrated by machines, as opposed to intelligence displayed by humans or by other animals. ...
## 2. History of artificial intelligence
Path: A/History_of_artificial_intelligence
Snippet: The history of artificial intelligence (AI) began in antiquity, with myths, stories and rumors of artificial beings endowed with intelligence or consciousness by master craftsmen. ...
## 3. Philosophy of artificial intelligence
Path: A/Philosophy_of_artificial_intelligence
Snippet: The philosophy of artificial intelligence is a branch of the philosophy of technology that explores artificial intelligence and its implications for knowledge, reality, consciousness, and the human mind. ...
Getting ZIM entries:
{
"name": "get_zim_entry",
"arguments": {
"zimFilePath": "D:/ZIM/wikipedia_en_all_nopic_2023-07.zim",
"entryPath": "A/Artificial_intelligence"
}
}
Response:
# Artificial intelligence
Path: A/Artificial_intelligence
Type: text/html
## Content
Artificial intelligence (AI) is intelligence demonstrated by machines, as opposed to intelligence displayed by humans or by other animals. Examples of specific artificial intelligence applications include expert systems, natural language processing, and computer vision.
Leading AI textbooks define the field as the study of "intelligent agents": any system that perceives its environment and takes actions that maximize its chance of achieving its goals. Some popular accounts use the term "artificial intelligence" to describe machines that mimic "cognitive" functions that humans associate with the human mind, such as "learning" and "problem solving".
...
License
MIT
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。