mcp-pve

[![npm version](https://img.shields.io/npm/v/@samik081/mcp-pve)](https://www.npmjs.com/package/@samik081/mcp-pve)
[![Docker image](https://ghcr-badge.egpl.dev/samik081/mcp-pve/latest_tag?trim=major&label=docker)](https://ghcr.io/samik081/mcp-pve)
[![License: MIT](https://img.shields.io/npm/l/@samik081/mcp-pve)](https://opensource.org/licenses/MIT)
[![Node.js Version](https://img.shields.io/node/v/@samik081/mcp-pve)](https://nodejs.org)

# MCP PVE

MCP server for [Proxmox VE](https://www.proxmox.com/en/proxmox-virtual-environment/overview). Manage virtual machines, containers, storage, networking, and clusters through natural language in Cursor, Claude Code, and Claude Desktop.

## Features

- **121 tools** across **12 categories** covering the Proxmox VE REST API
- **Three access tiers** (`read-only`, `read-execute`, `full`) for granular control
- **Category filtering** via `PVE_CATEGORIES` to expose only the tools you need
- **Zero HTTP dependencies** -- uses native `fetch` (Node 22+)
- **Self-signed cert support** via `PVE_VERIFY_SSL=false`
- **Docker images** for `linux/amd64` and `linux/arm64` on [GHCR](https://ghcr.io/samik081/mcp-pve)
- **Remote MCP** via HTTP transport (`MCP_TRANSPORT=http`) using the Streamable HTTP protocol
- **TypeScript/ESM** with full type safety

## API Compatibility

Tested with Proxmox VE **9.2.4**.

## Quick Start

Run the server directly with npx:

```bash
PVE_BASE_URL="https://pve.example.com:8006" \
PVE_TOKEN_ID="root@pam!mcp" \
PVE_TOKEN_SECRET="xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx" \
npx -y @samik081/mcp-pve
```

The server validates your PVE connection on startup and fails immediately with a clear error if credentials are missing or invalid.

### Docker

Run with Docker (stdio transport, same as npx):

```bash
docker run --rm -i \
  -e PVE_BASE_URL=https://pve.example.com:8006 \
  -e PVE_TOKEN_ID=root@pam!mcp \
  -e PVE_TOKEN_SECRET=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx \
  -e PVE_VERIFY_SSL=false \
  ghcr.io/samik081/mcp-pve
```

To run as a remote MCP server with HTTP transport:

```bash
docker run -d -p 3000:3000 \
  -e MCP_TRANSPORT=http \
  -e PVE_BASE_URL=https://pve.example.com:8006 \
  -e PVE_TOKEN_ID=root@pam!mcp \
  -e PVE_TOKEN_SECRET=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx \
  -e PVE_VERIFY_SSL=false \
  ghcr.io/samik081/mcp-pve
```

The MCP endpoint is available at `http://localhost:3000` and a health check at `http://localhost:3000/health`.

## Configuration

**Claude Code CLI (recommended):**

```bash
# Using npx
claude mcp add --transport stdio pve \
  --env PVE_BASE_URL=https://pve.example.com:8006 \
  --env PVE_TOKEN_ID=root@pam!mcp \
  --env PVE_TOKEN_SECRET=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx \
  --env PVE_VERIFY_SSL=false \
  -- npx -y @samik081/mcp-pve

# Using Docker
claude mcp add --transport stdio pve \
  --env PVE_BASE_URL=https://pve.example.com:8006 \
  --env PVE_TOKEN_ID=root@pam!mcp \
  --env PVE_TOKEN_SECRET=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx \
  --env PVE_VERIFY_SSL=false \
  -- docker run --rm -i ghcr.io/samik081/mcp-pve

# Using remote HTTP (connect to a running Docker container or HTTP server)
claude mcp add --transport http pve http://localhost:3000
```

**JSON config** (works with Claude Code `.mcp.json`, Claude Desktop `claude_desktop_config.json`, Cursor `.cursor/mcp.json`):

```json
{
  "mcpServers": {
    "pve": {
      "command": "npx",
      "args": ["-y", "@samik081/mcp-pve"],
      "env": {
        "PVE_BASE_URL": "https://pve.example.com:8006",
        "PVE_TOKEN_ID": "root@pam!mcp",
        "PVE_TOKEN_SECRET": "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
        "PVE_VERIFY_SSL": "false"
      }
    }
  }
}
```

**Docker (stdio):**

```json
{
  "mcpServers": {
    "pve": {
      "command": "docker",
      "args": ["run", "--rm", "-i",
        "-e", "PVE_BASE_URL=https://pve.example.com:8006",
        "-e", "PVE_TOKEN_ID=root@pam!mcp",
        "-e", "PVE_TOKEN_SECRET=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
        "-e", "PVE_VERIFY_SSL=false",
        "ghcr.io/samik081/mcp-pve"
      ]
    }
  }
}
```

**Remote MCP** (connect to a running Docker container or HTTP server):

```json
{
  "mcpServers": {
    "pve": {
      "type": "streamable-http",
      "url": "http://localhost:3000"
    }
  }
}
```

## Access Tiers

Control which tools are available using the `PVE_ACCESS_TIER` environment variable:

| Tier | Tools | Description |
|------|-------|-------------|
| `full` (default) | 121 | Read, execute, and write -- full control |
| `read-execute` | 76 | Read and execute -- no resource creation/deletion |
| `read-only` | 54 | Read only -- safe for exploration, no state changes |

**Tier details:**

- **full**: All 121 tools. Includes creating/deleting VMs, containers, storage, users, firewall rules, and more.
- **read-execute**: 76 tools. All read tools plus power actions (start, stop, migrate), bulk guest actions, OCI image pull, backup execution, and task management.
- **read-only**: 54 tools. List, get, status, and log tools only. No state changes.

Tools that are not available in your tier are not registered with the MCP server. They will not appear in your AI tool's tool list, keeping the context clean.

## Environment Variables

| Variable | Required | Default | Description |
|----------|----------|---------|-------------|
| `PVE_BASE_URL` | Yes | -- | URL of your PVE instance (e.g. `https://pve:8006`) |
| `PVE_TOKEN_ID` | Yes | -- | API token ID (`user@realm!tokenname`) |
| `PVE_TOKEN_SECRET` | Yes | -- | API token UUID secret |
| `PVE_ACCESS_TIER` | No | `full` | `read-only`, `read-execute`, or `full` |
| `PVE_CATEGORIES` | No | *(all)* | Comma-separated category allowlist |
| `PVE_TOOL_BLACKLIST` | No | *(none)* | Comma-separated list of tool names to exclude (e.g., `pve_delete_qemu_vm`) |
| `PVE_TOOL_WHITELIST` | No | *(none)* | Comma-separated list of tool names to force-include, bypassing access tier and category filters |
| `PVE_VERIFY_SSL` | No | `true` | Set `false` for self-signed certs |
| `DEBUG` | No | `false` | Enable debug logging to stderr |
| `MCP_TRANSPORT` | No | `stdio` | Transport mode: `stdio` (default) or `http` |
| `MCP_PORT` | No | `3000` | HTTP server port (only used when `MCP_TRANSPORT=http`) |
| `MCP_HOST` | No | `0.0.0.0` | HTTP server bind address (only used when `MCP_TRANSPORT=http`) |
| `MCP_SESSION_IDLE_TIMEOUT` | No | `1800` | Seconds of inactivity after which an HTTP session is closed and its memory released (sessions with a connected SSE stream are never closed). `0` disables. Only used when `MCP_TRANSPORT=http` |
| `MCP_EXCLUDE_TOOL_TITLES` | No | `false` | Set `true` to omit tool titles from registration (saves tokens) |

Create API tokens in the PVE UI under **Datacenter > Permissions > API Tokens**. Make sure to uncheck "Privilege Separation" if you want the token to inherit the user's full permissions.

### Available Categories

`nodes`, `qemu`, `lxc`, `storage`, `cluster`, `access`, `pools`, `network`, `firewall`, `backup`, `tasks`, `ha`

## Tools

mcp-pve provides 121 tools organized by category. Each tool's Access column shows the minimum tier required: `read-only` (available in all tiers), `read-execute` (requires `read-execute` or `full`), or `full` (requires `full` tier only). The Hints column shows tool behavior: `read-only` (no state changes), `destructive` (modifies existing state), `idempotent` (same result if called twice).

<details>
<summary>Nodes (8 tools)</summary>

| Tool | Description | Access | Hints |
|------|-------------|--------|-------|
| `pve_list_nodes` | List all nodes in the cluster | read-only | read-only, idempotent |
| `pve_get_node_status` | Get detailed node status (CPU, memory, uptime, load) | read-only | read-only, idempotent |
| `pve_get_node_version` | Get PVE version info for a node | read-only | read-only, idempotent |
| `pve_get_node_dns` | Get DNS settings for a node | read-only | read-only, idempotent |
| `pve_get_node_time` | Get time and timezone info for a node | read-only | read-only, idempotent |
| `pve_get_node_syslog` | Get system log entries from a node | read-only | read-only, idempotent |
| `pve_list_node_services` | List all system services on a node | read-only | read-only, idempotent |
| `pve_manage_node_service` | Start, stop, restart, or reload a node service | read-execute | destructive |

</details>

<details>
<summary>QEMU Virtual Machines (20 tools)</summary>

| Tool | Description | Access | Hints |
|------|-------------|--------|-------|
| `pve_list_qemu_vms` | List all QEMU VMs on a node | read-only | read-only, idempotent |
| `pve_get_qemu_status` | Get current VM status (CPU, memory, disk, network) | read-only | read-only, idempotent |
| `pve_get_qemu_config` | Get VM configuration | read-only | read-only, idempotent |
| `pve_get_qemu_rrddata` | Get RRD statistics over a time period | read-only | read-only, idempotent |
| `pve_list_qemu_snapshots` | List all VM snapshots | read-only | read-only, idempotent |
| `pve_start_qemu_vm` | Start a VM | read-execute | destructive |
| `pve_stop_qemu_vm` | Stop a VM (immediate) | read-execute | destructive |
| `pve_shutdown_qemu_vm` | Gracefully shut down a VM | read-execute | destructive |
| `pve_reboot_qemu_vm` | Reboot a VM | read-execute | destructive |
| `pve_suspend_qemu_vm` | Suspend a VM | read-execute | destructive |
| `pve_resume_qemu_vm` | Resume a suspended VM | read-execute | destructive |
| `pve_reset_qemu_vm` | Reset a VM (hard) | read-execute | destructive |
| `pve_migrate_qemu_vm` | Migrate a VM to another node | read-execute | destructive |
| `pve_create_qemu_vm` | Create a new VM | full | — |
| `pve_delete_qemu_vm` | Delete a VM and all its data | full | destructive |
| `pve_update_qemu_config` | Update VM configuration | full | destructive, idempotent |
| `pve_clone_qemu_vm` | Clone a VM | full | — |
| `pve_create_qemu_snapshot` | Create a VM snapshot | full | — |
| `pve_delete_qemu_snapshot` | Delete a VM snapshot | full | destructive |
| `pve_rollback_qemu_snapshot`