GPUtw Docs

AI assistant

Let Claude Code, Cursor, Codex or Copilot call the GPUtw API for you.

Install the skill

GPUtw publishes an open knowledge package that teaches an AI coding assistant this API — the endpoints, the auth header, how to poll a deploy until it is running, and how to clean up afterwards. Install it once and you can ask for an instance in plain language instead of writing the calls yourself.

Example
git clone https://github.com/GPUtw-ai/GPUtw-Skill.git ~/.claude/skills/gputw

That path is where Claude Code looks for skills. Restart the assistant and mention GPUtw in a prompt to pick it up.

What it knows

  • Authentication: which header your API key goes in, and which scope each call needs.
  • GPU inventory: reading the catalog and finding a machine with free capacity.
  • Instances: creating, polling to running, stopping and deleting — including the waits that are easy to get wrong.
  • Ports, vault and billing: exposing a service, writing to /vault so data survives a stop, and reading usage.

Other assistants

The same package works with Cursor, Codex CLI, GitHub Copilot and Gemini CLI, and ships an MCP server exposing the API as structured tools if your client speaks MCP. The repository README carries the per-client setup.

github.com/GPUtw-ai/GPUtw-Skill

Check what it writes

Warning

The skill is a separate repository on its own release cycle, so it can lag a change to this API. Treat generated code as a draft: read the call it produced against these docs before running it against a live account, and keep an eye on anything that stops or deletes an instance.