# Computegpu

> Computegpu is a paid MCP server for cloud and infrastructure. Rent cloud GPUs from your agent: browse the catalogue, deploy, stop, destroy. It can run locally on your machine or as a hosted remote server.

- **Kind:** MCP server
- **Pricing:** Paid
- **Platforms:** API, CLI, Web
- **Website:** https://computegpu.com
- **Repository:** https://github.com/jyswee/computegpu-cli
- **MCP Registry name:** io.github.jyswee/computegpu
- **Categories:** cloud and infrastructure, AI models and LLM tooling
- **Listing updated:** 2026-11-15

## About

Where most GPU clouds are driven from a dashboard, this one is built so an agent can provision hardware itself. Sign-up gives a gpu_ API key, and the owner verifies a card with a $0 charge before the first deploy. Wallet credit is topped up through Stripe from $1.

Instances are billed by the second with no hourly rounding, and a stopped instance stops billing so it can be resumed later. The site lists 22 GPU models, while the README mentions 47, with consumer cards such as the RTX 3060 at $0.22 an hour up to datacenter parts like the H200 at $10.76. Spot instances are 50% off but can be preempted with a five-minute warning.

Deploy options include multi-GPU counts, Docker with SSH, persistent volumes and pre-loaded Ollama models behind an Ollama-compatible endpoint. Clusters of 2 to 32 instances with InfiniBand are offered too. The CLI and a REST API cover the same lifecycle. The CLI is proprietary.

## Features

- **Per-second billing:** Instances are billed per second of uptime, with no hourly rounding.
- **Live GPU catalogue:** Lists current GPU models and rates, and finds the cheapest card for a VRAM target.
- **Spot pricing:** Interruptible instances at 50% off.
- **Ollama endpoint:** Deploys with pre-loaded models behind an Ollama-compatible API.
- **Pause and resume:** Stopping an instance halts billing and it can be resumed later.
- **Multi-node clusters:** Clusters of 2 to 32 instances with InfiniBand.

## Good for

- Let an agent spin up a GPU for a fine-tuning run
- Serve an Ollama model on a dedicated GPU
- Find the cheapest GPU with enough VRAM
- Run distributed training on a small cluster
- Pause a dev instance overnight

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http computegpu https://mcp.computegpu.com/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://mcp.computegpu.com/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "computegpu": {
      "url": "https://mcp.computegpu.com/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "computegpu": {
      "type": "http",
      "url": "https://mcp.computegpu.com/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "computegpu": {
      "serverUrl": "https://mcp.computegpu.com/mcp"
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "computegpu": {
      "type": "http",
      "url": "https://mcp.computegpu.com/mcp"
    }
  }
}
```

## FAQ

### How is it billed?

Per second of uptime, funded from a prepaid wallet with a $1 minimum top-up.

### Is a card required?

Yes, the owner verifies a card with no charge before the first deploy.

### What do spot instances cost?

50% off, with preemption after a five-minute warning.

### Is the CLI open source?

No. The README lists it as proprietary.

### Is Computegpu free?

No. Computegpu is a paid product.

### How do I add Computegpu to Claude Code?

Run this command in your terminal: claude mcp add --transport http computegpu https://mcp.computegpu.com/mcp

### What are the best alternatives to Computegpu?

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

## Alternatives

- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [Agent Yurt](https://ai-powered.com/mcp/agent-yurt): Agent Yurt is a free MCP server for knowledge and memory.
- [Proofline](https://ai-powered.com/mcp/proofline): Proofline is a freemium MCP server for developer tools.
- [signafAIde](https://ai-powered.com/mcp/signafaide): signafAIde is a freemium MCP server for security.
- [Toolkien](https://ai-powered.com/mcp/toolkien): Toolkien is an open-source MCP server for ai models and llm tooling.
- [filex](https://ai-powered.com/mcp/filex): filex is an open-source MCP server for files and documents.
- [Lugh](https://ai-powered.com/mcp/lugh): Lugh is an open-source MCP server for ai models and llm tooling.
- [HypePaper](https://ai-powered.com/mcp/hypepaper): HypePaper is a freemium MCP server for ai models and llm tooling.

---
Source: https://ai-powered.com/mcp/computegpu
