Getting Started

Two commands. One GPU. You're up and running in under a minute.

1

Install the CLI

Ozekai ships a lightweight CLI. Install it with pip:

pip install ozekai
Requirements: Python 3.10+ and pip. Works on Linux, macOS, and WSL.
2

Connect to your GPU

Run the connect command — it will prompt you for your username and password:

connect-gpu

That's it. You now have a dedicated GPU machine with CUDA, PyTorch, and JupyterLab running at http://localhost:8888.

3

Start working — Jupyter is already running

JupyterLab launches automatically when you connect. Just open your browser:

http://localhost:8888

No setup, no tokens, no passwords — click the link and you're in a notebook. Or run scripts directly from the terminal:

python train.py
4

Disconnect when done

Billing stops the instant you disconnect. No idle charges, ever.

disconnect-gpu

Or disconnect directly from your Dashboard with one click — no terminal needed.

CLI Reference

Three commands. That's the whole thing.

connect-gpu Connect to your GPU machine and start a billed session. Prompts for credentials. JupyterLab auto-starts.
disconnect-gpu Disconnect and immediately stop billing. Your work is saved. You can also do this from the Dashboard.
upload-gpu Upload files to your container. Supports glob patterns (*.py). See full syntax →

Don't have an account yet?

See Plans & Pricing