Quickstart

• A ClockingCloud account. Create one with email and password or Google.
• Prepaid credit. The minimum top-up is $10. See Billing.
• macOS or Linux with Homebrew. On Windows, use WSL 2.

brew install clockingcloud/tap/clockingcloud
clockingcloud version

The package installs the clockingcloud command. Every example in these docs uses it.

clockingcloud login
clockingcloud whoami

clockingcloud login opens your browser, you approve the request, and the CLI stores a credential locally. On a machine without a browser run clockingcloud login --device.

clockingcloud run z-image-turbo \
  --prompt "a ceramic mug on a linen table, soft window light" \
  --out ./shots

The CLI starts a GPU, loads the model, generates the image, writes it to ./shots, and releases the GPU. You pay for the seconds the GPU was busy. The first run of a model can take longer because its weights are cached on first use. Later runs are warm.

Create clockingcloud.yaml in your project. It is the same file you use locally, in review, and in CI/CD.

name: product-shots
model: z-image-turbo
gpu: auto
scale:
  min: 0
  max: 8
idle_timeout: 60s

setup:
  - pip install -r requirements.txt

volumes:
  - name: assets
    mount: /data
    mode: ro
  - name: renders
    mount: /out

clockingcloud plan
clockingcloud deploy -f clockingcloud.yaml
clockingcloud status product-shots

clockingcloud plan validates the file and prints the GPU and estimated cost without creating anything. With scale.min: 0 the instance costs nothing until the first request arrives.

clockingcloud balance
clockingcloud usage --since 1d
clockingcloud stop product-shots

Stopping releases the GPUs immediately. The definition and volumes remain, so clockingcloud start product-shots brings it back.

• Running models for every clockingcloud run option.
• Deploying instances for scaling, idle timeouts and CI/CD.
• Storage to connect S3, Azure Blob or Google Cloud Storage.
• Configuration reference for every clockingcloud.yaml field.