This post walks you through how to run an MCP (Model Context Protocol) server on LMStudio using powerful functions from the NetApp DataOps Toolkit. This setup lets you implement custom storage actions directly from your LLM-powered workspace on your laptop or workstation, enabling data management for AI and ML workflows.
For more information on MCP servers with the NetApp DataOps Toolkit, refer to this post.
Prerequisites
To run an MCP server in LMStudio for the NetApp DataOps Toolkit, you would need the following -
- Download and install LMStudio here. Ensure you have enough storage and memory to run a model locally.
- Install uv.
- Create a DataOps Toolkit config file using the command shown here.
Steps
Before you jump into the implementation, it is necessary to add or modify the mcp.json file in LMStudio. You can find it under 'Integrations'. Below is an example of mcp.json config file.
{
"mcpServers": {
"netapp_dataops_ontap_mcp": {
"command": "uvx",
"args": [
"--from",
"netapp-dataops-traditional",
"netapp_dataops_ontap_mcp.py"
]
},
"netapp_dataops_k8s_mcp": {
"command": "uvx",
"args": [
"--from",
"netapp-dataops-k8s",
"netapp_dataops_k8s_mcp.py"
]
}
}
}
The following video is a walkthrough of running an MCP server on LMStudio using NetApp DataOps Toolkit functions.
Using this minimal setup, you can implement an MCP server with the NetApp DataOps Toolkit and run it locally on your machine.
Refer to the NetApp DataOps Toolkit Github repository for more information.
