Rancher MCP Server

Model Context Protocol (MCP) server for the Rancher ecosystem: multi-cluster Kubernetes, Harvester HCI (VMs, storage, networks), and Fleet GitOps. Exposed over MCP by the rancher mcp server, that capability becomes something an assistant can invoke while it works, not something you go and do afterwards.

What it actually does

• Harvester toolset — List/get VMs, images, volumes, networks, hosts; VM actions; addon list/switch (enable/disable)
• Kubernetes toolset — List/get/create/patch/delete resources by apiVersion/kind; describe (resource + events), events, capacity
• Helm toolset — List/get/history of releases; install, upgrade, rollback, uninstall; repo list
• Fleet toolset — GitRepo list/get/create; Bundle list; Fleet cluster list; drift detection
• Rancher APIs — Same Bearer token for Steve (/k8s/clusters/...) and Norman (/v3/...); no CLI wrappers
• Security — Read-only default, disable-destructive, sensitive data masking (Norman token/credential fields redacted unless --show-sensitive-data)

Its toolset

Everything the assistant can do here goes through one of these:

• harvester_vm_list — List VMs with status, namespace, spec/status
• harvester_vm_get — Get one VM (full spec and status)
• harvester_vm_action — start, stop, restart, pause, unpause, migrate
• harvester_vm_create — Create VM (when not read-only). Supports network, interface_type (managedtap/bridge/masquerade), subnet for KubeOVN VPC
• harvester_vm_snapshot — Create/list/restore/delete VM snapshots
• harvester_vm_backup — Create/list/restore VM backups (Backup Target)
• harvester_image_list — List VM images (VirtualMachineImage)
• harvester_image_create — Create VM image from URL (when not read-only)
• harvester_volume_list — List PVCs (Longhorn-backed volumes)
• harvester_volume_create — Create volume/PVC (optionally from image)
• harvester_network_list — List VM networks (NetworkAttachmentDefinition)
• harvester_network_create — Create VM network - KubeOVN overlay or VLAN (when not read-only)

Adding it to your client

rancher-mcp-server on npm is all you need. Most clients run it directly, so configuration is a few lines and a restart.

Configuration

You will need 5 environment variables: RANCHER_MCP_RANCHER_SERVER_URL, RANCHER_MCP_RANCHER_TOKEN, RANCHER_MCP_TOOLSETS, YOUR_TOKEN, YOUR_RANCHER_URL. The server will not start without them, which is usually why the tools fail to appear on a first run. Keep credentials in your client's env block or a secrets manager rather than in a file you might commit.

When to reach for it

Plenty of cloud and infrastructure servers cover similar ground. The differences that matter in practice are scope of access and how much setup stands between you and a working tool call. Rancher's toolset — harvester_vm_list, harvester_vm_get, harvester_vm_action and 11 more — is a fair guide to whether it matches your workflow. It is maintained by mrostamii; worth a glance at recent repository activity before you build anything load-bearing on it.

SyncDev reviews every entry in this directory against the project's own documentation before publishing, and revisits them as servers change.

Caveats

• It runs with your machine's permissions. That is convenient and also the reason to think about what you point it at before you approve a tool call.
• With 14 tools registered it takes up a noticeable share of the context window; turn it off in projects that never touch Rancher.
• Missing credentials fail quietly in some clients — if no tools show up, check the environment block first.
• MCP clients confirm each tool call by default. Leave that on until you have watched what the rancher mcp server does with a few real requests.