OPS
@nanovms
OPS is a MCP server that allows you to build, deploy and orchestrate unikernels on any cloud from AWS to GCP or Azure. Unikernels are single application virtual machines with no users, no shells and no ability to run other programs.
Overview
What is OPS?
OPS is an MCP server for ops (operations) tasks. It provides tools to list instances, list images, and create instances via the Model Context Protocol.
How to use OPS?
Build the server normally, then configure it in claude_desktop_config.json with the correct path to the ops-mcp binary and necessary environment variables (HOME, LOGNAME, PATH, SHELL, USER). Use the tools List instances, List images, or Instance create <image_name> (e.g., Instance create redis-server).
Key features of OPS
- Lists available instances.
- Lists available images.
- Creates new instances from specified image names.
Use cases of OPS
- Retrieve a list of running instances managed by Ops.
- View available images for deployment.
- Quickly launch a new instance (e.g., redis-server) from an image.
- Automate instance management through Claude conversations.
FAQ from OPS
What tools does OPS provide?
List instances, List images, and Instance create <image_name>.
How is OPS configured?
Add an entry to claude_desktop_config.json specifying the path to the ops-mcp binary and environment variables such as HOME, LOGNAME, PATH, SHELL, and USER.
What are the runtime requirements?
The README notes to ensure the command is in the correct path and the PATH variable is set appropriately. The server is built “like normally”, implying a local Go build environment.
Is OPS production-ready?
The README states it is a very initial cut and the author is open to suggestions, indicating early development.
