Public API.

GPU.ai exposes a JSON REST API for everything you can do in the dashboard: list GPU types, provision instances, manage SSH keys and webhooks, and pull usage data. Authenticate with a Bearer API key, point at https://api.demo.gpu.ai/v1, and you're live.

§ 01.1Get an API key¶

Keys have the shape gpuai_live_<24 chars> and are shown exactly once at creation. We only store a SHA-256 hash. There is no recovery if you lose the raw key.

Option A

From the dashboard

Sign in, open API Keys in the cloud dashboard, and click Create key. Copy the raw key from the modal before closing it.

Option B

From the CLI

Install the gpu CLI and run gpu login. The device flow mints a key and writes it to ~/.config/gpu/credentials.json with mode 0600.

§ 01.2Hello, world (no auth)¶

The catalog endpoints are public: no API key required. List every GPU type in the marketplace:

curl https://api.demo.gpu.ai/v1/gpu-types

Responses are JSON envelopes with a data array and a next_cursor field. See the pagination conventions.

§ 01.3Provision your first instance¶

Writes need a Bearer key and an Idempotency-Key header. Instance creation is async: the response is 202 Accepted with an Operation-Id you can poll at GET /v1/operations/{id}.

curl -X POST https://api.demo.gpu.ai/v1/instances \
  -H "Authorization: Bearer gpuai_live_..." \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "gpu_type": "h100_sxm",
    "gpu_count": 1,
    "tier": "on_demand",
    "ssh_key_ids": ["sshkey_01HX..."]
  }'

§ 01.4Let your coding agent do this¶

The gpu CLI ships a skill that teaches coding agents (Claude Code and Codex) to rent and drive GPU.ai instances themselves: quote prices, provision, SSH in, run the workload, and clean up. One command installs it for every agent on your machine:

SHELL

gpu skill install
export GPUAI_API_KEY=gpuai_live_...   # agents can't do the browser login flow

Then ask your agent to "spin up an H100 and run train.py on it". The skill includes cost-safety rules: the agent quotes the $/hr price before creating anything and deletes instances when the work is done. See the CLI guide for details.

§ 01.5OpenAPI spec¶

The full OpenAPI 3.1 spec is published at github.com/gpuai-dev/openapi and served live at api.demo.gpu.ai/v1/openapi.json. Pipe it into Postman, Stoplight, or any OpenAPI tool to explore interactively.