> For the complete documentation index, see [llms.txt](https://submodel.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://submodel.gitbook.io/docs/overview.md).

# Welcome

Submodel.ai provides container-based compute management, queued Serverless jobs, and access to model services. Choose between compute you deploy yourself and inference APIs supplied by the separate model service.

These docs were aligned with the complete backend `2699f35cf8f04b57a491003a6e7401a89b6d48e7` and webapp `a3fe1688d216a8ea12721953f2dfb752dc094938` snapshots. Plans, stock, pricing, quotas, and deployment configuration affect availability. These snapshots do not fix node-client, Python SDK, model-gateway, or container-image versions.

## Choose a service

| Need                                                           | Entry                                                                             | Workflow                                                                                                                                |
| -------------------------------------------------------------- | --------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------- |
| Interactive development, training, or a persistent application | [Pods](https://submodel.gitbook.io/docs/pod/overview.md)                          | Select a plan and template/image, manage the application inside the container, and connect or inspect logs from details                 |
| Request-driven work with adjustable Worker counts              | [Serverless](https://submodel.gitbook.io/docs/serverless/overview.md)             | Deploy an image implementing job claiming and result callbacks, submit jobs to the instance Endpoint, and query status                  |
| Hosted model inference                                         | [Model service / InstaGen](https://submodel.gitbook.io/docs/instagen/overview.md) | The console currently links to an external site; platform metadata and Key management are separate from the inference gateway           |
| Connect your own hardware                                      | [Hosting](https://submodel.gitbook.io/docs/hosting/overview.md)                   | Obtain the client from the connection page, register the device, and complete environment checks, readiness, and resource configuration |

## Pods

An instance stores configuration and target state; running Pods host the actual containers. GPUs per plan and Pods per instance are separate quantities. The public deployment page exposes On-Demand (`payg` in the API) and Spot. Scheduling also depends on the zone, stock, storage location, and capacity protection.

Successful creation still requires scheduling, image preparation, and application startup. Stop, restart, reset, and release have different effects on data. Read [manage Pods](https://submodel.gitbook.io/docs/pod/manage-pods.md) and [storage types](https://submodel.gitbook.io/docs/pod/types.md) before operating on an instance.

## Serverless

An Endpoint is identified by its instance ID. Jobs enter through `run`; Workers claim, process, and report results; clients query `status`. Minimum/maximum Workers, idle timeout, and scaling policy control resource targets without guaranteeing fixed cold-start latency or job concurrency.

Charges measure running Worker time, including time kept running while idle, rather than only job execution time. vLLM images need a platform adapter; see [OpenAI compatibility](https://submodel.gitbook.io/docs/serverless/overview-2/overview/openai-compatibility.md) for the supported surface.

## Get started

1. [Register and deploy your first instance](https://submodel.gitbook.io/docs/getting-started/get-started.md).
2. Read [billing, manual recharge, and balance alerts](https://submodel.gitbook.io/docs/getting-started/billing-information.md).
3. Before automating resource management, review [platform API Keys](https://submodel.gitbook.io/docs/getting-started/api-keys.md) and [SDK compatibility](https://submodel.gitbook.io/docs/sdk/overview.md).
4. Check current [Pod plans](https://submodel.ai/#/plan/pod) and [Serverless configuration](https://submodel.ai/#/plan/serverless) in the console; confirm model availability and prices on the actual model-service page.

For help, provide the instance ID, time, and redacted error information to [support](https://submodel.gitbook.io/docs/references/contact-us.md).
