# Center AI Vision GPU cloud

> Center AI Vision GPU cloud is a paid MCP server for cloud and infrastructure. MCP server to rent A100 and H100 GPUs by the minute from your agent within a spending cap. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Paid
- **Platforms:** Web, API, CLI
- **Website:** https://console.centeraivision.com/agents
- **MCP Registry name:** com.centeraivision/gpu-cloud
- **Categories:** cloud and infrastructure, AI models and LLM tooling
- **Listing updated:** 2026-11-15

## About

Center AI Vision's server lets a coding agent find a GPU, start it, set it up over SSH or Jupyter and stop it, all against your prepaid balance. It shows the price and asks before starting or extending a machine, and you can cap spending with an API key that has a spending limit and expiry; the key can never commit more GPU time than that limit.

The workflow is quote, then start: quote_gpu prices one or more GPUs for some hours and holds the price for 10 minutes, checking that a named model fits for vLLM, and start_gpu takes the total from your balance. Setups include PyTorch with Jupyter, vLLM for serving an LLM, and Ubuntu with SSH. Stopping a machine returns unused time to your balance.

Machines stop on their own when paid hours end, and files on a machine are deleted when it stops, so results must be copied off first. A connected app or key can rent and stop GPUs but cannot change billing, team or keys. A public endpoint, /mcp/public, lists GPUs, setups and estimated prices without sign-in.

## Features

- **Spending-limited keys:** Keys carry a spending cap and expiry so unattended agents cannot overspend.
- **Quote then start:** Prices are held for 10 minutes and nothing is charged until start_gpu.
- **Three setups:** PyTorch with Jupyter, vLLM for LLM serving, or Ubuntu with SSH.
- **Per-minute rental:** Stopping a machine returns unused time to your balance.
- **Public price lookup:** A no-sign-in endpoint lists GPUs, setups and estimated prices.
- **Activity trail:** Actions appear in account activity marked with the key or app used.

## Good for

- Train a model on an A100 for a few hours and stop the machine afterwards
- Serve an open LLM across two H100s
- Check current GPU prices and availability
- Run unattended CI jobs with a capped key

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http gpu-cloud https://console.centeraivision.com/mcp
```

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

```sh
https://console.centeraivision.com/mcp
```

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

```json
{
  "mcpServers": {
    "gpu-cloud": {
      "url": "https://console.centeraivision.com/mcp"
    }
  }
}
```

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

```json
{
  "servers": {
    "gpu-cloud": {
      "type": "http",
      "url": "https://console.centeraivision.com/mcp"
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "gpu-cloud": {
      "serverUrl": "https://console.centeraivision.com/mcp"
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "gpu-cloud": {
      "type": "http",
      "url": "https://console.centeraivision.com/mcp"
    }
  }
}
```

## Tools

- `list_gpus`: Every GPU with hourly price and free availability
- `list_setups`: PyTorch + Jupyter, vLLM, or Ubuntu + SSH
- `get_account`: Balance, ID check, running machines and remaining key limit
- `quote_gpu`: Price GPUs for some hours and hold the price for 10 minutes
- `start_gpu`: Start the quoted GPU and take the total from balance
- `list_machines`: Machines starting or running with cost and connection info
- `wait_for_machine`: Wait for a new machine to be ready
- `extend_machine`: Add hours to a running machine
- `stop_machine`: Stop a machine and return unused time
- `list_ssh_keys`: List saved SSH keys
- `add_ssh_key`: Save a public key for SSH machines
- `set_machine_ssh_keys`: Change which keys can log in to a running machine

## FAQ

### Can an agent overspend?

A key can never commit more GPU time than its spending limit, and the agent shows the price and asks before starting or extending.

### What happens to my files?

Files on a machine are deleted when it stops, so copy results off first.

### Can I try it without an account?

Yes. The public server lists GPUs, setups and estimated prices; starting a GPU needs the full server.

### Can the agent change billing?

No. A connected app or key cannot change billing, team or keys.

### How do I add funds?

On the website, under Billing.

### Is Center AI Vision GPU cloud free?

No. Center AI Vision GPU cloud is a paid product.

### How do I add Center AI Vision GPU cloud to Claude Code?

Run this command in your terminal: claude mcp add --transport http gpu-cloud https://console.centeraivision.com/mcp

### What are the best alternatives to Center AI Vision GPU cloud?

The closest alternatives to Center AI Vision GPU cloud 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/center-ai-vision-gpu-cloud
