Skip to main content

Overview

The Massed Compute MCP server lets AI assistants like Claude, Cursor, Codex, and ChatGPT interact with your Massed Compute account on your behalf — listing GPU inventory, launching and managing instances, auditing billing, and following setup recipes.

No SDK, no custom code. Your AI client speaks the Model Context Protocol, and our server translates. Authenticate with an API token or an OAuth 2.1 grant (the same header either way).

What you can do with it

  • "What's running right now?" Ask your assistant to list your instances, regions, and hourly cost without leaving your editor.
  • "Spin up an H100 for this notebook." Have Claude or Cursor pick a GPU matching your workload, deploy it, wait for boot, and hand you the SSH command.
  • "What's the cheapest way to fine-tune Llama 3 70B?" Get a recommendation grounded in our live inventory and pricing — no manual table-comparing.

Server details

FieldValue
Endpointhttps://vm.massedcompute.com/api/mcp
AuthAuthorization: Bearer <credential> on every JSON-RPC POST
API tokenFrom API Settings — user_uuid.api_key
OAuth tokenIssued after consent — mcat_… access tokens. Review grants under Connected Apps

Clients that speak MCP OAuth discover it from a 401 + WWW-Authenticate on this endpoint. You do not paste a token in that path — you sign in and approve access.

Before you connect

API token path

  1. Create a key at vm.massedcompute.com/settings/api.
  2. Pick read-only when the assistant should only look at fleet, billing, and SSH keys. Pick full when it should launch or terminate.

OAuth path

  1. Point a compatible client at https://vm.massedcompute.com/api/mcp with no Bearer header.
  2. Complete the Massed Compute sign-in and consent screen (full access vs read-only).
  3. Manage or revoke the app later under Connected Apps.

Next steps

  • Set up your client — copy-pastable config for Claude Code, Cursor, Codex, and Claude Desktop.
  • Tools reference — what your AI assistant can do, with example prompts.
  • Agent Skills — downloadable workflow guides for GPU selection, multi-VM launches, cost audits, and safe cleanup.
  • Troubleshooting — common issues and fixes.