# aide-mcp-server [Health: Active]

**Category:** 📊 Data Platforms  
**Repository:** https://github.com/NetApp/aide-mcp-server  
**GitHub Stars:** 0  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/aide-mcp-server

## Description
MCP server for NetApp AI Data Engine

## Claude Desktop Quick Installation
Remote MCP endpoint (confidence: high). Install path detected from listing signals. Add as a URL/SSE server in your client:

```json
"mcpServers": {
  "aide-mcp-server": {
    "url": "https://lmstudio.ai/docs/app/mcp"
  }
}
```

## Documentation

## What aide-mcp-server MCP server does

The aide-mcp-server MCP server exposes NetApp AI Data Engine's retrieval-augmented generation search capability to MCP-compatible clients. Its primary function is document retrieval: an agent submits a natural-language query, and the server passes it to the AIDE RAG API, which uses vector-based semantic similarity to find relevant documents.

The repository package is named `netapp-aide-mcp`, and the exposed MCP tool is `netapp_data_engine_search`. This makes the server suitable for assistants that need to ground responses in content managed by NetApp AI Data Engine rather than relying only on a model's built-in knowledge. It belongs to the **🔎 end to end RAG platforms** category because it provides the MCP interface to a hosted RAG search system.

## How it works

Communication uses the stdio transport, so the server runs locally as a process started by the MCP client. It is not presented as a hosted HTTP MCP endpoint. Most supported desktop clients start the process from their MCP configuration.

Authentication is deferred until the first tool call. With PKCE authentication, the flow opens a browser window. With device code authentication, the server prints a short code and verification details to its console logs; the user completes the flow on a device with browser access. Only these two authentication flows are supported.

## Setup and configuration

The aide-mcp-server MCP server requires Python 3.10 or newer and `uv`. Before launching it, create a `.netapp` JSON file in the user's home directory. The file needs the AIDE RAG API endpoint, SSL verification behavior, the selected authentication flow, and the corresponding identity-provider endpoints and parameters.

PKCE configuration includes a redirect URI and enables PKCE in the token request parameters. Device code configuration instead specifies a device-code endpoint and the client scope. The repository includes an example configuration file in its `Examples` directory. Protect the file from access by other users; on Unix-like systems, `chmod 600 ~/.netapp` is the documented approach.

Run the published package with `uvx --from netapp-aide-mcp server`, or point `uvx` at a cloned repository path to run from source. Add the command as a stdio server in the MCP client's JSON configuration.

## Tools and capabilities

The aide-mcp-server MCP server provides:

- `netapp_data_engine_search` for searching documents through the AIDE RAG API.
- Vector-based semantic similarity retrieval driven by a query.
- PKCE authentication for machines with browser access.
- Device code authentication for environments without a local browser.

## Limitations and notes

This server provides search access rather than a general-purpose set of NetApp management operations. The supplied material documents one MCP tool only. It also depends on a correctly formatted `.netapp` file, valid AIDE and identity-provider settings, Python 3.10+, and `uv`.

Because the transport is stdio, clients must be able to run local stdio MCP servers. Hosted clients that do not support stdio may not work. Authentication does not begin when the process starts; troubleshooting should focus on the first tool call and, for device code flow, the server console logs. If a client does not invoke the tool, the README suggests explicitly mentioning NetApp AIDE in the prompt.

_Full upstream README: https://allmcps.com/mcp/aide-mcp-server/readme_