Skip to content
MCP server

CloudGPU MCP

Lets a coding assistant rent a per-minute GPU, run a job and shut it down

Runs locallyStatus and history
Pricing
Paid
Platforms
CLI, API
Runs
Locally
Version
0.12.1
Listing updated
3 days ago

About CloudGPU MCP

CloudGPU MCP is a paid MCP server for cloud and infrastructure. It runs locally on your machine and is started by your AI client.

The pitch is a hand-off: tell the assistant to fine-tune something on a 4090 and bring back the adapter. It quotes a price, rents the GPU only after you agree, uploads code and data, runs the job, downloads results and stops the machine.

The server runs on your own computer, so the assistant can reach the repository and datasets you already have open. Budgets and time limits are enforced by CloudGPU's servers on every instance, so a GPU still shuts down if the chat ends or the laptop sleeps. The docs list per-minute and whole-hour options; per-minute options refund unused prepaid minutes when you destroy the instance, while whole-hour ones charge each started hour in full.

Node.js 18 or newer is required.

Install

Run in a terminal

claude mcp add cloudgpu-mcp --env CLOUDGPU_API_KEY='<CLOUDGPU_API_KEY>' -- npx -y cloudgpu-mcp

Add to claude_desktop_config.json

{
  "mcpServers": {
    "cloudgpu-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "cloudgpu-mcp"
      ],
      "env": {
        "CLOUDGPU_API_KEY": "<CLOUDGPU_API_KEY>"
      }
    }
  }
}

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "cloudgpu-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "cloudgpu-mcp"
      ],
      "env": {
        "CLOUDGPU_API_KEY": "<CLOUDGPU_API_KEY>"
      }
    }
  }
}

Add to .vscode/mcp.json

{
  "servers": {
    "cloudgpu-mcp": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "cloudgpu-mcp"
      ],
      "env": {
        "CLOUDGPU_API_KEY": "<CLOUDGPU_API_KEY>"
      }
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "cloudgpu-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "cloudgpu-mcp"
      ],
      "env": {
        "CLOUDGPU_API_KEY": "<CLOUDGPU_API_KEY>"
      }
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "cloudgpu-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "cloudgpu-mcp"
      ],
      "env": {
        "CLOUDGPU_API_KEY": "<CLOUDGPU_API_KEY>"
      }
    }
  }
}

Configuration

  • CLOUDGPU_API_KEYOptional CloudGPU API key. Without a key, use public discovery and connect_account for an owner-approved scoped key.

Tools this server provides (1)

Taken from the project documentation; the running server may differ.

name
description

What it does

Quote before renting
Shows a price estimate and rents only after approval.
Budget and time limits
Limits are enforced server-side on each rented instance.
Run jobs and commands
Runs commands or jobs and tails logs on the instance.
Download results
Brings results back before the GPU is stopped.
Scoped keys
connect_account lets the owner approve a scoped key instead of pasting one.
One balance
The same key and balance cover GPUs and the model API.

Good for

  • Fine-tune a model on a rented 4090 from your editor
  • Run a one-off training job without managing a cloud account
  • Cap an agent's GPU spend with a budget
  • Let an agent pick a GPU from a public catalog

Pricing

Paid

Prepaid balance, billed per minute or per started hour; prices not listed here

Prices change. Check cloudgpu.app before you buy.

CloudGPU MCP FAQ

What Node version is needed?

Node.js 18 or newer.

What if my laptop sleeps?

Server-side limits still shut the instance down.

Is the API key required?

It is optional; the assistant can call connect_account to get an owner-approved scoped key.

Are unused minutes refunded?

On per-minute options, yes when you destroy the instance; whole-hour options are not.

Is CloudGPU MCP free?

No. CloudGPU MCP is a paid product.

How do I add CloudGPU MCP to Claude Code?

Run this command in your terminal: claude mcp add cloudgpu-mcp --env CLOUDGPU_API_KEY='<CLOUDGPU_API_KEY>' -- npx -y cloudgpu-mcp

What are the best alternatives to CloudGPU MCP?

The closest alternatives to CloudGPU MCP are Claude Code Ultimate Guide, Agent Yurt, Proofline, signafAIde and Toolkien.

Alternatives to CloudGPU MCP

  1. MCP server for searching the Claude Code Ultimate Guide from any MCP client

    Runs locally
    • Open source
    • 6.5K GitHub stars
    • AI models and LLM tooling
  2. Agent Yurt

    MCP server

    Back up agent memory, files and dashboards, then recover from a reviewed snapshot

    Not checked yet
    • Remote
    • Free
    • Knowledge and memory
  3. Proofline

    MCP server

    Production-aware review of changes against your AWS or Google Cloud setup, via MCP

    • Remote
    • Freemium
    • Developer tools
  4. signafAIde

    MCP server

    Local MCP server to set up, verify and check coverage of signed agent action records

    Runs locally
    • 5 tools
    • Freemium
    • Security
  5. Toolkien

    MCP server

    One MCP gateway for all your MCP servers, with code mode, permissions and audit

    Not checked yet
    • 3 tools
    • Remote
    • Open source
    • AI models and LLM tooling
  6. filex

    MCP server

    Self-hosted file manager in one Go binary with a built-in MCP server for agents

    Not checked yet
    • Remote
    • Open source
    • Files and documents