> For the complete documentation index, see [llms.txt](https://docs.nebulablock.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.nebulablock.com/getting-started/get-started/products.md).

# Products

What Nebula Block offers — serverless inference, GPU cloud, and object storage — and which tier each one needs.

Nebula Block runs three product lines that share one account, one credit balance, and one API key.

## Serverless Inference

Managed endpoints for state-of-the-art models — text, multimodal chat, vision, image generation, video generation, embeddings, and reranking. Nothing to provision: authenticate and call.

* **Endpoint:** `https://inference.nebulablock.com/v1` (OpenAI-compatible)
* **Billing:** pay-as-you-go, metered per token (per image or per second for some media models)
* **Access:** all tiers, but per-model daily caps apply — see [Tiers and Rate Limits](/account/tiers-and-limits.md)

→ [Serverless Inference](/products/serverless-inference.md) · [Model Catalog](/products/serverless-inference/model-catalog.md)

## GPU Cloud

On-demand GPU and CPU instances billed by the hour. Pick your hardware and OS image, attach an SSH key, and the instance is reachable in minutes. Instances can be started, stopped, rebooted, and firewalled from the console or the Platform API.

* **Access:** CPU instances from Tier 2, GPU instances from Tier 3
* **Billing:** hourly, deducted from your credit balance for as long as the instance exists — powered on or not

→ [GPU Cloud](/products/gpu-cloud.md) · [Quickstart](/products/gpu-cloud/quickstart.md)

> **Note:** Instances are deleted automatically if your credit balance runs out. Keep auto-pay on, or watch your balance, for anything long-running.

## Object Storage

S3-compatible object storage for datasets, checkpoints, and generated media. Works with `s3cmd`, `boto3`, and the AWS SDKs for Go and Java, and can be managed through the console or the Platform API.

→ [Object Storage](/products/object-storage.md) · [Quickstart](/products/object-storage/quickstart.md)

## Also available in the console

These are live in the [console](https://console.nebulablock.com) but not yet covered in depth by these docs:

|                          | What it is                                                            |
| ------------------------ | --------------------------------------------------------------------- |
| **Reserved Instances**   | Committed-term GPU capacity at a lower effective rate than on-demand  |
| **Dedicated Endpoints**  | Single-tenant inference deployments                                   |
| **Hardware Store**       | Purchase used and refurbished server hardware — GPUs, storage, memory |
| **Sovereign AI Compute** | Canadian-hosted compute and models for data-residency requirements    |

## See also

* [Quickstart](/getting-started/get-started/quickstart.md)
* [Billing Information](/getting-started/get-started/billing.md)
* [Tiers and Rate Limits](/account/tiers-and-limits.md)