Runcrate

Biarkan agen Anda menyewa GPU cloud: luncurkan instans H100/A100/RTX 4090, jalankan perintah melalui SSH, panggil 200+ model, lacak pengeluaran. Remote, OAuth.

Server MCP Terhosting

npx add-mcp 'https://mcp.runcrate.ai/mcp'

Terpasang ke Claude Code, Codex, Cursor, dan lainnya

Dokumentasi

Setup

Connect Runcrate to Claude Code, Codex, Cursor, VS Code, Claude, ChatGPT or any MCP client: one line pasted into your agent, or one click.

Connect your agent

Paste this one line into your agent: Claude Code, Codex, Cursor, VS Code, OpenCode, or any agent that can run a command. It reads runcrate.ai/mcp.md, adds the server to itself, and its first Runcrate tool call opens your browser to sign in. There is no key to copy.

Add Runcrate's MCP server to yourself: run curl -fsSL "https://runcrate.ai/mcp.md" and follow it.

Add to Cursor

One click. Cursor asks to install, then opens your browser to sign in on first use.

Install in VS Code

One click. VS Code adds the server, then signs you in on first use.

The Claude Code plugin adds the same server and a skill on keeping spend in check. The Gemini CLI extension adds the server. Run either in a terminal:

claude plugin marketplace add runcrateai/runcrate-mcp && claude plugin install runcrate@runcrate
gemini extensions install https://github.com/runcrateai/runcrate-mcp

Claude and ChatGPT on the web

They can't run a command, so they take the URL as a custom connector: https://mcp.runcrate.ai/mcp.

In Claude (web and desktop):

  1. In Claude, open Settings, then Connectors, and choose Add custom connector.
  2. Name it Runcrate and paste the URL above.
  3. Choose Connect and sign in to Runcrate in the window that opens.

On Team and Enterprise plans, an owner adds the connector under organization settings first; each member then connects with their own account.

In ChatGPT (chatgpt.com/plugins):

  1. In ChatGPT, open Plugins, choose +, then Add custom MCP server.
  2. Name it Runcrate, paste the URL above and pick OAuth.
  3. Create it, sign in to Runcrate, then mention @Runcrate in a chat.

Set it up by hand

Claude Code:

claude mcp add --transport http --scope user runcrate https://mcp.runcrate.ai/mcp

Then run /mcp in Claude Code and pick runcrate: your browser opens to sign in.

Codex:

codex mcp add runcrate --url https://mcp.runcrate.ai/mcp

Codex opens your browser to sign in as it adds the server. Closed it? Run codex mcp login runcrate.

VS Code:

code --add-mcp '{"name":"runcrate","type":"http","url":"https://mcp.runcrate.ai/mcp"}'

Every client on this machine at once, with the Runcrate CLI (0.5.0 or later):

runcrate mcp install

It finds Claude Code, Codex, Cursor, VS Code, Windsurf and Gemini CLI, and connects each one. It backs up each config file before changing it, adds Runcrate next to your other servers, and never replaces an entry you already have. Pick clients with --client cursor,vscode, or preview with --dry-run.

Windsurf, Zed, Warp, Goose, Raycast, Devin and any other MCP client: give the client the URL https://mcp.runcrate.ai/mcp and choose OAuth. It signs you in through your browser on first use.

Cursor (~/.cursor/mcp.json), Warp (~/.warp/.mcp.json) and most clients:

{
  "mcpServers": {
    "runcrate": {
      "url": "https://mcp.runcrate.ai/mcp"
    }
  }
}

Claude Code project config (.mcp.json) needs the type:

{
  "mcpServers": {
    "runcrate": {
      "type": "http",
      "url": "https://mcp.runcrate.ai/mcp"
    }
  }
}

VS Code (mcp.json in your user profile or .vscode/mcp.json):

{
  "servers": {
    "runcrate": {
      "type": "http",
      "url": "https://mcp.runcrate.ai/mcp"
    }
  }
}

Windsurf (mcp_config.json): {"mcpServers":{"runcrate":{"serverUrl":"https://mcp.runcrate.ai/mcp"}}}

Gemini CLI (~/.gemini/settings.json): {"mcpServers":{"runcrate":{"httpUrl":"https://mcp.runcrate.ai/mcp"}}}

Zed (settings.json): "context_servers": {"runcrate":{"url":"https://mcp.runcrate.ai/mcp"}}

Codex (~/.codex/config.toml):

[mcp_servers.runcrate]
url = "https://mcp.runcrate.ai/mcp"

Authentication

The MCP server uses OAuth with your Runcrate account. When you first connect:

  1. Your AI assistant prompts you to authorize
  2. A browser window opens to the Runcrate login page
  3. You pick the workspace and the environment the agent works in, then approve
  4. The token is stored by your MCP client, so you won't need to re-authorize unless it expires

The agent then acts only in that environment, and that environment's monthly budget caps what it can spend. It can't create, delete or switch environments, or change a budget. To move it to another environment, disconnect Runcrate in your client and sign in again.

[!TIP]

For one-off work, ask your agent to run it as a job (run_job): the machine stops itself, and its billing, when the command ends. For an instance it keeps, ask for a max runtime and an idle stop. See Spend guardrails.

Rate Limiting

The MCP server allows 100 requests per minute per IP address. If you exceed this, requests return an error until the window resets.

This documentation is built and hosted on Mintlify, a developer documentation platform.