fabiante/gridscale-mcp
If you are the rightful owner of gridscale-mcp and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
This is a Model Context Protocol (MCP) server designed to facilitate AI models' interaction with the Gridscale API, enabling infrastructure provisioning through tools like 5ire.
create_ip
Tool to create a new IP address in Gridscale.
delete_ip
Tool to delete an existing IP address in Gridscale.
create_storage
Tool to create storage, with partial parameter support.
Gridscale MCP
This is a MCP Server which allows AI modells to interact with the Gridscale API
You can use this with tools such as 5ire to let your LLM provision infrastructre as you wish.
[!CAUTION] If you use this program, you are responsible for the consequences. You should only use this in an empty Gridscale project where the AI model will not cause any harm like deleting your production database.
Features (Tools and Resources)
With this MCP server you can use the following tools and resources:
| Type | Name | Note | Supported | Support Wanted |
|---|---|---|---|---|
| Tool | create_ip | Yes | ||
| Tool | delete_ip | Yes | ||
| Tool | create_storage | Not all parameters are supported | Partially | |
| Tool | create_server | No | Yes | |
| Tool | create_network | No | Yes | |
| Res | get_storage_templates | Yes |
If you want to contribute, try adding one of the above missing tools or resources.
Building + Installation
To build this program you need a Go environment.
The recommended way is to use make:
make build
This will have created a binary called gridscale-mcp in the project directory.
Install
It simplifies things if you install the binary to a location in your $PATH, for example /usr/bin on Linux.
You can use make for that:
# Install
make install
# Uninstall
make uninstall
Usage
[!CAUTION] If you use this program, you are responsible for the consequences. You should only use this in an empty Gridscale project where the AI model will not cause any harm like deleting your production database.
Prerequisites
You need:
- To have a build of
gridscale-mcp. Follow the steps in the Building section to first build the program. - Public API credentials for the Gridscale API
Usage in an MCP client
To run this server you need two arguments:
gridscale-mcp --user-key <your-gs-user-key> --user-token <your-gs-user-token>
I recommend to use 5ire as desktop client to interact with an LLM and this MCP server. You are responsible for setting up an LLM (for example Gemini).
In 5ire the MCP server is setup as tool like so:
{
"name": "Gridscale",
"key": "gridscale",
"command": "gridscale-mcp",
"args": [
"--user-key",
"...",
"--user-token",
"..."
]
}
Debugging
You can directly communicate with you MCP server as with any other API server.
The MCP Inspector is a great tool for this.
If you have Node.js installed, simply run:
make inspect