Proxmox VE MCP Server
MCP server giving AI assistants control over Proxmox VE VMs, containers and clusters
- Pricing
- Open source
- Platforms
- CLI, API, self-hosted
- Runs
- Locally
- Version
- 0.11.0
- Listing updated
- today
Similar MCP servers
About Proxmox VE MCP Server
Proxmox VE MCP Server is an open-source MCP server for cloud and infrastructure. It runs locally on your machine and is started by your AI client.
The README says this server connects Proxmox Virtual Environment to AI clients such as Cursor, Claude Code and Claude Desktop, so virtual machines, containers, storage, networking and clusters can be queried and changed in plain language. It covers roughly 129 tools grouped into 12 areas that follow the Proxmox REST API, including node diagnostics, QEMU and LXC lifecycle actions, snapshots, backups, high availability and access control.
Access is governed by a tier setting (read-only, read-execute or full) and by optional category and tool allowlists, so an operator can expose only monitoring tools or hide destructive operations. Authentication uses a Proxmox API token supplied through environment variables. It runs over stdio via npx, Docker or the npm package, and can also serve Streamable HTTP, which the README notes has no built-in authentication.
It suits Proxmox administrators and homelab operators who want an assistant to check node health, review task logs or start and migrate guests. The README is explicit that the HTTP mode should stay on a trusted network or behind an authenticating proxy.
Install
Run in a terminal
claude mcp add mcp-pve --env PVE_BASE_URL='<PVE_BASE_URL>' --env PVE_TOKEN_ID='<PVE_TOKEN_ID>' --env PVE_TOKEN_SECRET='<PVE_TOKEN_SECRET>' -- npx -y @samik081/mcp-pveAdd to claude_desktop_config.json
{
"mcpServers": {
"mcp-pve": {
"command": "npx",
"args": [
"-y",
"@samik081/mcp-pve"
],
"env": {
"PVE_BASE_URL": "<PVE_BASE_URL>",
"PVE_TOKEN_ID": "<PVE_TOKEN_ID>",
"PVE_TOKEN_SECRET": "<PVE_TOKEN_SECRET>"
}
}
}
}Add to ~/.cursor/mcp.json
{
"mcpServers": {
"mcp-pve": {
"command": "npx",
"args": [
"-y",
"@samik081/mcp-pve"
],
"env": {
"PVE_BASE_URL": "<PVE_BASE_URL>",
"PVE_TOKEN_ID": "<PVE_TOKEN_ID>",
"PVE_TOKEN_SECRET": "<PVE_TOKEN_SECRET>"
}
}
}
}Add to .vscode/mcp.json
{
"servers": {
"mcp-pve": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@samik081/mcp-pve"
],
"env": {
"PVE_BASE_URL": "<PVE_BASE_URL>",
"PVE_TOKEN_ID": "<PVE_TOKEN_ID>",
"PVE_TOKEN_SECRET": "<PVE_TOKEN_SECRET>"
}
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"mcp-pve": {
"command": "npx",
"args": [
"-y",
"@samik081/mcp-pve"
],
"env": {
"PVE_BASE_URL": "<PVE_BASE_URL>",
"PVE_TOKEN_ID": "<PVE_TOKEN_ID>",
"PVE_TOKEN_SECRET": "<PVE_TOKEN_SECRET>"
}
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"mcp-pve": {
"command": "npx",
"args": [
"-y",
"@samik081/mcp-pve"
],
"env": {
"PVE_BASE_URL": "<PVE_BASE_URL>",
"PVE_TOKEN_ID": "<PVE_TOKEN_ID>",
"PVE_TOKEN_SECRET": "<PVE_TOKEN_SECRET>"
}
}
}
}Using several servers? Build one config file with the MCP config generator.
Configuration
PVE_BASE_URLURL of the Proxmox VE instance (e.g. https://pve.example.com:8006)RequiredPVE_TOKEN_IDProxmox VE API token ID (e.g. user@pam!token-name)RequiredPVE_TOKEN_SECRETProxmox VE API token secretRequiredPVE_BASE_URLURL of the Proxmox VE instance (e.g. https://pve.example.com:8006)RequiredPVE_TOKEN_IDProxmox VE API token ID (e.g. user@pam!token-name)RequiredPVE_TOKEN_SECRETProxmox VE API token secretRequired
What it does
- Tiered access control
- Read-only, read-execute and full tiers decide which tools are registered with the client.
- Category and tool filtering
- Environment variables allow limiting exposure to chosen tool categories or specific tool names.
- Broad Proxmox API coverage
- Tools span nodes, QEMU VMs, LXC containers, storage, cluster, firewall, backup, HA and access control.
- Compact metrics output
- RRD statistics tools return bucketed tables instead of large raw sample sets.
- Optional HTTP transport
- Can run as a remote Streamable HTTP server alongside the default stdio mode.
- Self-signed certificate support
- A setting disables TLS verification for PVE instances using self-signed certificates.
Good for
- Checking node health, disk SMART status and running services on a Proxmox cluster
- Starting, stopping or migrating VMs and containers from an AI assistant
- Creating snapshots before upgrades and reviewing backup recency
- Reviewing task logs and cluster HA status during troubleshooting
- Running read-only monitoring of a Proxmox environment without change rights
Pricing
Open source
Free and MIT-licensed; self-hosted.
Proxmox VE MCP Server FAQ
Which Proxmox VE version is it tested with?
The README says it was tested with Proxmox VE 9.2.4.
What is needed to connect it?
A PVE base URL and an API token ID and secret are required; the server validates the connection on startup.
How can I stop the assistant from making changes?
Setting the access tier to read-only registers 62 read tools only, and read-execute adds power actions but no creation or deletion.
Is the HTTP transport secured?
The README states the HTTP transport has no authentication and recommends loopback, a trusted network, or an authenticating reverse proxy.
Is Proxmox VE MCP Server free?
Yes. Proxmox VE MCP Server is open source, so you can use and self-host it for free.
How do I add Proxmox VE MCP Server to Claude Code?
Run this command in your terminal: claude mcp add mcp-pve --env PVE_BASE_URL='<PVE_BASE_URL>' --env PVE_TOKEN_ID='<PVE_TOKEN_ID>' --env PVE_TOKEN_SECRET='<PVE_TOKEN_SECRET>' -- npx -y @samik081/mcp-pve
What are the best alternatives to Proxmox VE MCP Server?
The closest alternatives to Proxmox VE MCP Server are Proxmox VE, Komodo MCP Server, Proxmox VE (Mojood Connect), VPS MCP & Digital Secretary and rust-proxmoxmcp.
Where this comes from
- Description: written from the project's README and its MCP Registry entry.
- Install steps: from the official MCP Registry entry.