This feature is in Public Preview.
The MCP connector is available in the following Nerdio Manager plans:
|
AVD Core |
|
|
AVD Premium |
|
|
Windows 365 |
|
|
Unified Endpoint Management |
|
Deploying the MCP connector requires the Admin role in Nerdio Manager. Once deployed, any user with read access to a given module can use the MCP connection to access data for that functional area.
Some AI agents can't perform the OAuth sign-in flow themselves. For these agents, authentication is instead handled via a Nerdio Manager bridging app registration, specified in the MCP configuration file.
Deploying the MCP connector has no impact on Nerdio Manager performance at scale, and no specific preparation is required. However, because the MCP connection queries live data, your AI agent may take several minutes to respond when connected to larger environments.
-
You need access to a supported AI client with MCP connectivity.
-
Node.js must be available on the client machine running your AI tool. Some AI tools (e.g. Visual Studio Code) include this by default; others require it to be installed separately.
-
In Nerdio Manager, go to System > Settings > Integrations > Model Context Protocol.
-
Toggle the Status and Nerdio Manager Permissions | Read switches to Enabled.
-
Use the dropdown to select the client type you want to connect to Nerdio Manager via the MCP connector.
A configuration code snippet appears for you to copy into your AI client's MCP configuration page.
Nerdio Manager provides a configuration snippet for each supported client type. Copy the snippet for the AI tool you're using direct from the configuration menu in Nerdio Manager (see above). The format for each supported client type is shown for reference below.
Visual Studio Code
{
"servers": {
"nerdio-mcp": {
"url": "https://nmw-app-<Your-Instance>.azurewebsites.net/mcp",
"type": "http"
}
}
}
Codex
[mcp_servers.nerdio-mcp]
command = "npx"
args = [
"-y",
"@nerdio-inc/mcp-bridge",
"https://nmw-app-<Your-Instance>.azurewebsites.net/mcp"
]
Other client AI tools
{
"mcpServers": {
"nerdio-mcp": {
"command": "npx",
"args": [
"-y",
"@nerdio-inc/mcp-bridge",
"https://nmw-app-<Your-Instance>.azurewebsites.net/mcp"
]
}
}
}
Replace <Your-Instance> with the URI from your own NME instance.
-
On the client machine, open the AI tool that will connect to the MCP server and go to its MCP settings. This will typically be located at a menu path like Settings > Developer > Local MCP servers > Edit config.
-
Add the copied snippet to your AI tool's configuration file (JSON/TOML) and save.
-
Fully close the AI tool and relaunch it. On load, you're prompted to authorize access for the tool to your Nerdio environment through Azure, in your browser.
-
Once authorized, the MCP server configuration is available to view in the AI tool's settings.
-
The first time you use the MCP server in a query, you're prompted to allow your AI tool to use this MCP connection.
Once you accept the prompt, your AI client is configured to use the MCP connector.
The following are example prompts you can use once your AI agent is connected to the Nerdio Manager MCP server, with screenshots showing example output.
Use the Nerdio MCP server to show me all my current host pools in my NME environment
Can you show a summary of the auto-scaling applied to the DFG-Multisession host pool?
Which tenants is my NME instance connected to?
Do I need to update my NME instance?
How many users are currently connected to my AVD environment?
Can you use the MCP server to write me a PowerShell script that will deploy the latest Windows 11 updates to my images?
Contact our Sales team for more information about this feature.
Raise a support ticket for this item.
Comments (0 comments)