NetApp AIDE MCP Server

NetApp AIDE MCP Server

MCP server that exposes NetApp AI Data Engine's RAG search for semantic document retrieval.

Category
访问服务器

README

NetApp AI Data Engine (AIDE) MCP Server

Description

netapp-aide-mcp is an MCP server (Python package) that exposes the NetApp AI Data Engine's RAG (Retrieval-Augmented Generation) search functionality via MCP. The server exposes a tool named netapp_data_engine_search, which provides the ability to search for documents using AIDE's RAG API. This RAG API implements a vector-based semantic similarity search engine that retrieves relevant documents based on the provided query.

[!NOTE] This MCP server uses the stdio transport, making it a "local MCP server".

Quick start

Prerequisites

  • Python >= 3.10
  • uv (manages all installations automatically)

Configuration

Before running the server, you need to create a .netapp file in your home directory with the necessary configuration.

  1. Create the .netapp file:

    • Open a terminal or file explorer.
    • Navigate to your home directory (e.g., ~ on Unix-like systems or C:\Users\YourUsername on Windows).
    • Create a new file named .netapp.
  2. Add the JSON configuration:

    • Open the .netapp file in a text editor.
    • Add the following JSON configuration, replacing the example values with your own:

    For PKCE flow: (Recommended if you have a browser available on your machine)

    {
      "rag_search_api_endpoint_url": "https://example.com/api",
      "verify_ssl": true,
      "auth_flow": "pkce",
      "token_request_endpoint_url": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/authorize",
      "token_exchange_endpoint_url": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token",
      "token_request_params": {
        "client_id": "your_client_id",
        "redirect_uri": "http://localhost:8888",
        "scope": "api://your-app/.default",
        "use_pkce": true,
        "auth_timeout_seconds": 300
      }
    }
    

    For device code flow: (Use this if you do not have a browser on your machine. A short code will be printed in the logs — copy it, open the provided verification URL on any device, and enter the code to complete authentication.)

    {
      "rag_search_api_endpoint_url": "https://example.com/api",
      "verify_ssl": true,
      "auth_flow": "device_code",
      "device_code_endpoint_url": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/devicecode",
      "token_request_endpoint_url": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token",
      "token_request_params": {
        "client_id": "your_client_id",
        "scope": "api://your-app/.default"
      }
    }
    

    Only PKCE (web-based) and Device Code flows are supported.

  3. Set file permissions:

    • Ensure that the .netapp file is not readable by other users/groups for security reasons.

    • On Unix-like systems, you can set the permissions using the following command:

      chmod 600 ~/.netapp
      
    • On Windows, you can set the file permissions through the file properties dialog.

[!TIP] There is an Examples folder in the repository that contains a .netapp.example file. This file provides examples of how your .netapp file should look. You can use this as a reference when creating your own .netapp file.

Running with uvx

You can run the MCP server instantly, without installing anything globally.

[!NOTE] Authentication is initiated on the first tool call, not at server startup. When the first tool call is initiated: if you are using the PKCE flow, a browser window will open; if you are using the device code flow, the device code details will be printed to the MCP server's console logs.

Run pre-built package from PyPI

# Run the latest stable version
uvx --from netapp-aide-mcp server

# Run a specific version
uvx --from netapp-aide-mcp==1.0.0 server
  • server script launches the MCP server

Build and run from source

# Clone the repo
git clone https://github.com/NetApp/aide-mcp-server
cd aide-mcp-server

# Optional: check out a specific release
# git checkout tags/release-v1.0.0

# Retrieve the directory path for the cloned repo
export AIDE_MCP_REPO_PATH=$(pwd)

uvx --from $AIDE_MCP_REPO_PATH server
  • server script launches the MCP server

Example JSON config

To use this MCP server with an MCP client, you need to configure the client to use this server. For many clients (such as VS Code, Claude Desktop, and LMStudio), this requires editing a config file in JSON format (often named mcp.json). Below is an example. Refer to the documentation for your MCP client for specific formatting details.

[!NOTE] Most MCP clients will take care of starting the MCP server. You typically do not need to start the server yourself. As always, it's best to refer to the documentation for your MCP client for details on how it handles MCP servers.

{
	"servers": {
		"netapp-aide-mcp": {
			"type": "stdio",
			"command": "uvx",
			"args": [
				"--from",
				"netapp-aide-mcp",
				"server"
			]
		}
	}
}

Troubleshooting

  • Ensure your .netapp file is present and correctly formatted.
  • Check that Python 3.10+ and uv are installed.
  • Ensure that your MCP client is compatible with the stdio transport; most desktop clients (e.g. VS Code, Claude Desktop, LMStudio) are, but some hosted clients are not.
  • If you are using the device code auth flow, check the MCP server console logs for the device code details. The device code details will be printed to the logs when the first tool call is initiated.
  • If your client is not invoking an MCP tool, try adding "Be sure to use NetApp AIDE" to your prompt.

MCP Registry metadata

mcp-name: io.github.NetApp/aide-mcp-server

License

Distributed under the terms of the BSD 3-Clause License (see the LICENSE file in the repository).

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选