> See the full index of documentation pages at https://docs.zettabyte.com/llms.txt.

# Introduction

> Build, train, and serve AI models on GPUs.

Source: https://docs.zettabyte.com/docs/get-started/introduction

Last updated: 2026-07-21

zCLOUD is GPU compute for developing, training, fine-tuning, and serving AI models.

## Start here

- Quickstart: from sign-in to a deployed endpoint.
- Open the console: launch GPUs and deploy models.
- API reference: create keys and call your endpoint.

## What you can do

- Run on GPUs: launch and manage instances.
- Develop in notebooks: `Jupyter` or `VSCode` on a GPU.
- Train and fine-tune: training and inference jobs.
- Bring models and data: models, datasets, images.
- Deploy and serve: serve models as endpoints.
- Store your files: volumes and S3 storage.

## Two areas of the console

The console has two areas: a build area for instances, notebooks, jobs, models, data, and storage, and a serve area for inference endpoints, API keys, and token usage. Both share the same account, organization, and billing. The console labels these areas zCLOUD and zSTUDIO.